U.S. patent application number 16/731533 was filed with the patent office on 2020-12-31 for systems and methods for preparing compositions.
This patent application is currently assigned to Colgate-Palmolive Company. The applicant listed for this patent is Colgate-Palmolive Company. Invention is credited to Dong Hyun KIM, Bartosz LUCZYNSKI, Iraklis PAPPAS, Donghui WU.
Application Number | 20200411140 16/731533 |
Document ID | / |
Family ID | 1000005272717 |
Filed Date | 2020-12-31 |
United States Patent
Application |
20200411140 |
Kind Code |
A1 |
PAPPAS; Iraklis ; et
al. |
December 31, 2020 |
SYSTEMS AND METHODS FOR PREPARING COMPOSITIONS
Abstract
A system, apparatus, and/or method is disclosed for producing a
personal care product. Values of chemoinformatic properties of
ingredients of a sample chemical composition are received. A value
of a property of the sample chemical composition is received in
which the property is affected by an interaction of at least two of
the ingredients. The values of the chemoinformatic properties of
the ingredients of the sample chemical composition and the value of
the property of the sample chemical composition are input into a
model. An identity of the considered chemical composition is
determined, via the model, based on at least one of (1) values of
chemoinformatic properties of ingredients of the considered
chemical composition or (2) a value of a property of the considered
chemical composition. The personal care product comprised of the
considered chemical composition is produced.
Inventors: |
PAPPAS; Iraklis;
(Pennsauken, NJ) ; LUCZYNSKI; Bartosz; (Dover,
NJ) ; WU; Donghui; (Bridgewater, NJ) ; KIM;
Dong Hyun; (Somerset, NJ) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Colgate-Palmolive Company |
New York |
NY |
US |
|
|
Assignee: |
Colgate-Palmolive Company
New York
NY
|
Family ID: |
1000005272717 |
Appl. No.: |
16/731533 |
Filed: |
December 31, 2019 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
16672922 |
Nov 4, 2019 |
10839941 |
|
|
16731533 |
|
|
|
|
16452214 |
Jun 25, 2019 |
10515715 |
|
|
16672922 |
|
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G16C 20/10 20190201;
G16C 20/70 20190201; G16C 20/30 20190201 |
International
Class: |
G16C 20/30 20060101
G16C020/30; G16C 20/70 20060101 G16C020/70; G16C 20/10 20060101
G16C020/10 |
Claims
1. A method comprising: (a) receiving values of chemoinformatic
properties of ingredients of a sample chemical composition; (b)
receiving a value of a property of the sample chemical composition,
the property being affected by an interaction of at least two of
the ingredients; (c) inputting, into a model, the values of the
chemoinformatic properties of the ingredients of the sample
chemical composition and the value of the property of the sample
chemical composition; (d) determining, via the model, an identity
of the considered chemical composition based on at least one of (1)
values of chemoinformatic properties of ingredients of the
considered chemical composition or (2) a value of a property of the
considered chemical composition; and (e) producing a personal care
product comprised of the considered chemical composition. wherein
the property of the considered chemical composition is affected by
an interaction of at least two of the ingredients of the considered
chemical composition; and wherein steps (c)-(d) are performed by
one or more processors.
2. The method of claim 1, wherein the determining of the identity
of the considered chemical composition is based on the values of
the chemoinformatic properties of the ingredients of the considered
chemical composition and not based on the value of the property of
the considered chemical composition.
3. The method of claim 1, wherein, for each of the sample chemical
composition and the considered chemical composition, each of the
ingredients has a respective chemoinformatic property.
4. The method of claim 1, further comprising inputting, into the
model, an identity of the sample chemical composition.
5. The method of claim 1, wherein the model is a machine learning
model relating to a supervised learning approach.
6. The method of claim 1, wherein the property of the considered
chemical composition is identical to the property of the sample
chemical composition.
7. The method of claim 1, wherein the value of the property of the
sample chemical composition is determined via an experimental
measurement of the sample chemical composition.
8. The method of claim 1, wherein the value of the property of the
sample chemical composition is comprised of at least one of a first
value of the property determined via an experimental measurement of
the sample chemical composition or a second value of the property
via a thermodynamic calculation.
9. The method of claim 1, wherein the personal care product is a
toothpaste.
10. The method of claim 1, comprising: repeating steps (a)-(c) for
a plurality of sample chemical compositions prior to performing
step (d).
11. A method for determining a value of a considered chemical
composition having ingredients, the method comprising: (a)
receiving an identity of a sample chemical composition having a
defined value of a chemical property, the sample chemical
composition being comprised of ingredients that are different than
ingredients of the considered chemical composition; (b) generating
a training set comprising the identity of the sample chemical
composition and the defined value of the chemical property of the
sample chemical composition; (c) constructing, based on the
training set, a model for determining a value of a chemical
property of the considered chemical composition; (d) determining,
via the model, the value of the chemical property of the considered
chemical composition based on an identity of the considered
chemical composition and the training set; (e) receiving, from the
model, the value of the chemical property of the considered
chemical composition; and (f) producing a personal care product
comprised of the considered chemical composition wherein steps
(c)-(e) are performed by one or more processors.
12. The method of claim 11, wherein, for each of the sample
chemical composition and the considered chemical composition, the
identity comprises values of chemoinformatic properties of each of
the ingredients of each chemical composition.
13. The method of claim 11, wherein, for each of the sample
chemical composition and the considered chemical composition, the
identity comprises identities of the ingredients of each chemical
composition.
14. The method of claim 11, wherein the model is a machine learning
model, the value of the value of the chemical property of the
considered chemical composition being determined via rules of the
machine learning model.
15. The method of claim 14, wherein the machine learning model
comprises a supervised learning approach.
16. The method of claim 15, wherein the supervised learning
approach comprises at least one of a Decision Tree rule, a Random
Forest rule, a Support Vector Machine rule, a Naive Bayes
Classification rule, or a Logistic Regression rule.
17. The method of claim 11, wherein the chemical property of the
considered chemical composition relates to a physiochemical
property of the considered chemical composition, the physiochemical
property relating to a physical property or a chemical property of
the considered chemical composition.
18. The method of claim 11, wherein the defined chemical property
of the sample chemical composition is identified via at least one
of an experimental measurement of the sample chemical composition
or a thermodynamic calculation of the sample chemical
composition.
19. The method of claim 11, wherein the defined chemical property
of the sample chemical composition is comprised of at least one of
a first chemical property identified via an experimental
measurement of the sample chemical composition or a second chemical
property identified via a thermodynamic calculation of the sample
chemical composition.
20. The method of claim 11, comprising: repeating steps (a)-(c) for
a plurality of chemical compositions prior to performing steps
(d)-(e).
Description
CROSS-REFERENCE TO RELATED PATENT APPLICATION
[0001] The present application is a continuation application of
U.S. Non-Provisional patent application Ser. No. 16/452,214, filed
Jun. 25, 2019, the contents of which are hereby incorporated herein
by reference in their entirety.
BACKGROUND
[0002] Many products are formed of chemical compositions. Chemical
compositions typically include many different ingredients. Each of
the ingredients have a particular value, such as a chemoinformatic
value, associated with the ingredient. Further, one or more
properties (e.g., a pH, a consumer perception) of the chemical
composition may have unique values, for example, based on the
ingredients within the composition. The value of the property of
the chemical composition may change due to the interaction of the
ingredients within the composition.
[0003] Conventional methods exist for predicting values of one or
more properties of a chemical composition. However, such methods
are often time consuming and nontrivial. For example, conventional
approaches to predicting a value of a pH of a chemical composition
include (1) experimentally measuring the chemical composition to
determine the pH value; and (2) performing a mathematical
prediction calculation to determine the pH value (e.g., using known
acidity constants, such as pKa values). These approaches, however,
are deficient due to the time and/or complexity involved in the
respective approaches. Thus, a system and/or method is desired that
can determine a value of a property of a chemical composition in a
way that requires less time and/or less complexity.
BRIEF SUMMARY
[0004] A system, apparatus, and/or method is disclosed for
producing a personal care product. Values of chemoinformatic
properties of ingredients of a sample chemical composition are
received. A value of a property of the sample chemical composition
is received in which the property is affected by an interaction of
at least two of the ingredients. The values of the chemoinformatic
properties of the ingredients of the sample chemical composition
and the value of the property of the sample chemical composition
are input into a model. An identity of the considered chemical
composition is determined, via the model, based on at least one of
(1) values of chemoinformatic properties of ingredients of the
considered chemical composition or (2) a value of a property of the
considered chemical composition. The personal care product
comprised of the considered chemical composition is produced.
[0005] A system, apparatus, and/or method is disclosed for
determining a value of a property of a considered chemical
composition. An identity of a sample chemical composition may be
received. A sample chemical composition may comprise ingredients.
Each of the ingredients may be associated with a value of a
chemoinformatic property of chemoinformatic properties of the
sample chemical composition. A value of a property of the sample
chemical composition may be received. The property of the sample
chemical composition may be affected by an interaction of at least
two of the ingredients of the sample chemical composition. The
value of the property of the sample chemical composition and at
least one of (1) the identity of the sample chemical composition or
(2) the values of the chemoinformatic properties of the ingredients
of the sample chemical composition may be input into a model. The
value of the property of the considered chemical composition may be
determined, via the model, based on at least one of (1) an identity
of the considered chemical composition or (2) values of
chemoinformatic properties of ingredients of the considered
chemical composition. The property of the considered chemical
composition may be affected by an interaction of at least two of
the ingredients of the considered chemical composition.
[0006] In another aspect, a value of a property of a considered
chemical composition may be determined. An identity of a sample
chemical composition may be received. The identify may include
ingredients. One or more (e.g., each) of the ingredients may be
associated with a value of a chemoinformatic property of
chemoinformatic properties of the sample chemical composition. A
value of a sample physiochemical property of the sample chemical
composition may be received. The sample physiochemical property of
the sample chemical composition may be affected by an interaction
of at least two of the ingredients of the sample chemical
composition. The value of the sample physiochemical property of the
sample chemical composition may be input into a model. The identity
of the sample chemical composition and/or the values of the
chemoinformatic properties of the ingredients of the sample
chemical composition may be input into a model. The value of a
considered physiochemical property of the considered chemical
composition may be determined via the model. The value may be based
on an identity of the considered chemical composition and/or values
of chemoinformatic properties of ingredients of the considered
chemical composition. The considered physiochemical property of the
considered chemical composition may be affected by an interaction
of at least two of the ingredients of the considered chemical
composition. The considered physiochemical property may be
different than the sample physiochemical property.
[0007] In another aspect, a value of a property of a considered
chemical composition may be determined. An identity of a
characteristic of interest may be received. Identities of sample
chemical compositions may be received. Each of the sample chemical
compositions may comprise ingredients each being associated with a
value of a chemoinformatic property of chemoinformatic properties
of the sample chemical composition. For each of the sample chemical
compositions, a value of a property that is affected by an
interaction of at least two of the ingredients of the sample
chemical composition may be received. For each (e.g., only each) of
the sample chemical compositions having the characteristic of
interest, the value of the property of the sample chemical
composition and at least one of (1) the identity of the sample
chemical composition or (2) the values of the chemoinformatic
properties of the ingredients of the sample chemical composition
may be input into a model. The value of the property of the
considered chemical composition may be determined via the model.
The value may be based on at least one of (1) an identity of the
considered chemical composition or (2) values of chemoinformatic
properties of the ingredients of the considered chemical
composition. The property of the considered chemical composition
may be affected by an interaction of at least two ingredients of
the considered chemical composition.
[0008] In another aspect, a product comprised of a considered
chemical composition having a value of a considered chemical
property may be produced. An identity of a sample chemical
composition comprising ingredients may be received. Each of the
ingredients may be associated with a value of a chemoinformatic
property of chemoinformatic properties of the sample chemical
composition. A value of a training chemical property of the sample
chemical composition may be received. The value of the training
chemical property may be based on at least one of an experimental
measurement of the value of the training chemical property or a
mathematical measurement of the value of the training chemical
property. A learning model may be constructed using values of the
chemoinformatic properties of the sample chemical composition and
the value of the training chemical property of the sample chemical
composition. The value of the considered chemical property may be
input into the learning model. Values of chemoinformatic properties
of the considered chemical composition having the value of the
considered chemical property may be determined via the learning
model and/or based on the value of the considered chemical
property. The product comprising the considered chemical
composition having the value of the considered chemical property
may be produced.
[0009] In another aspect, a value of a considered chemical
composition having ingredients may be determined. An identity of a
sample chemical composition having a defined value of a chemical
property may be received. The sample chemical composition may be
comprised of ingredients that may be different than ingredients of
the considered chemical composition. A training set comprising the
identity of the sample chemical composition and the defined value
of the chemical property of the sample chemical composition may be
generated. A model for determining a value of a chemical property
of the considered chemical composition may be constructed based on
the training set. The value of the chemical property of the
considered chemical composition may be determined via the model,
for example, based on an identity of the considered chemical
composition and the training set. The value of the chemical
property of the considered chemical composition may be
received.
[0010] In another aspect, an identity of a sample chemical
composition comprising ingredients may be received. Each of the
ingredients may be associated with a value of a chemoinformatic
property of chemoinformatic properties of the sample chemical
composition. A value of a fitting parameter associated with a value
of a property of the sample chemical composition may be received.
The property of the sample chemical composition may be affected by
an interaction of at least two of the ingredients of the sample
chemical composition. The fitting parameter value associated with
the value of the property of the sample chemical composition and/or
at least one of (1) the identity of the sample chemical composition
or (2) the values of the chemoinformatic properties of the
ingredients of the sample chemical composition may be input into a
model. A fitting parameter value of the considered chemical
composition may be determined via the model. The fitting parameter
value may be based on at least one of (1) an identity of the
considered chemical composition or (2) values of chemoinformatic
properties of ingredients of the considered chemical
composition.
[0011] In another aspect, a considered chemical composition may be
identified. Values of chemoinformatic properties of ingredients of
a sample chemical composition may be received. A value of a
property of the sample chemical composition may be received. The
property may be affected by an interaction of at least two of the
ingredients. The values of the chemoinformatic properties of the
ingredients of the sample chemical composition and the value of the
property of the sample chemical composition may be input into a
model. An identity of the considered chemical composition may be
determined via the model, for example, based on at least one of (1)
values of chemoinformatic properties of ingredients of the
considered chemical composition or (2) a value of a property of the
considered chemical composition. The property of the considered
chemical composition may be affected by an interaction of at least
two of the ingredients of the considered chemical composition.
[0012] In another aspect, identities of components of a first
composition may be received. Each of the components may have a
value of a predefined characteristic of predefined characteristics.
A value of a property of the first composition may be received. The
property may be affected by an interaction of at least two of the
components of the first composition. A learning model may be
trained using the values of the predefined characteristics of the
components of the first composition and the value of the property
of the first composition. Identities of second components of a
second composition may be provided to the learning model. At least
one of the second components may be different than at least one of
the first components. A value of a property of the second
composition may be determined via the learning model. The property
of the second composition may be affected by an interaction of at
least two of the components of the second composition.
[0013] In another aspect, a value of a property of a considered
chemical composition may be determined. An identity of the
considered chemical composition from chemical compositions
associated with a personal care product may be received. The
considered chemical composition may include ingredients. Values of
chemoinformatic properties may be received. Each value may be
associated with a respective one of the ingredients of the
considered chemical composition. The value of the property of the
considered chemical composition may be determined based on at least
one of (1) the identity of the considered chemical composition or
(2) the values of the chemoinformatic properties associated with
the respective one of the ingredients of the considered chemical
composition. The model may be trained by at least one of (1) an
identity of a chemical composition or (2) values of chemoinformatic
properties of ingredients of the chemical composition with a value
of a property of the chemical composition. The value of the
property of the considered chemical composition may be affected by
an interaction of at least two of the ingredients of the considered
chemical composition.
[0014] In another aspect, a model may be created to determine a
value of a property of a considered chemical composition. An
identity of a sample chemical composition comprising ingredients
may be received. Each of the ingredients may be associated with a
value of a chemoinformatic property of chemoinformatic properties
of the sample chemical composition. A value of a property of the
sample composition may be received. The property may be affected by
an interaction of at least two of the ingredients of the sample
chemical composition. A model may be trained to determine the value
of the property of the considered chemical composition by
processing the value of the property of the sample chemical
composition and at least one of (1) an identity of the sample
composition or (2) the values of the chemoinformatic properties of
the ingredients of the sample chemical composition. The model may
be configured to determine the value of the property of the
considered chemical composition based on at least one of (1) an
identity of the considered chemical composition or (2) values of
chemoinformatic properties of ingredients of the considered
chemical composition.
BRIEF DESCRIPTION OF THE DRAWINGS
[0015] The present invention will become more fully understood from
the detailed description and the accompanying drawings,
wherein:
[0016] FIG. 1A is a table of ingredients of an example
composition;
[0017] FIG. 1B is a table of example properties of composition in
which consumers may have a perception;
[0018] FIG. 2A is a block diagram of example components of a
composition, the components including ingredients and substances of
the composition;
[0019] FIG. 2B is a table of ingredients of another example
composition, providing identities of the ingredients and
percentages of the ingredients;
[0020] FIG. 3A is an example process for determining a value of a
composition using machine learning rules;
[0021] FIG. 3B is an example system for determining a value of a
composition;
[0022] FIG. 4 is another example process for determining a value of
a composition using machine learning rules;
[0023] FIG. 5 is a block diagram of an example system including a
user device;
[0024] FIG. 6 is a block diagram of an example system including a
training of a property engine;
[0025] FIG. 7 is a table of example functions of ingredients of a
composition;
[0026] FIG. 8 is a table of example classifications of ingredients
of a composition;
[0027] FIGS. 9A, 9B, 9C are block diagrams of an example training
of a machine learning model and receiving values from the machine
learning model;
[0028] FIGS. 10A, 10B, 10C are example graphical user interfaces
(GUIs) for training a property engine;
[0029] FIGS. 11A, 11B, 11C, 11D are example graphical user
interfaces (GUIs) for receiving a determined value via a property
engine; and
[0030] FIG. 12 is an example method of determining values of a
composition, as described herein.
DETAILED DESCRIPTION
[0031] The following description of the preferred embodiment(s) is
merely exemplary in nature and is in no way intended to limit the
invention or inventions. The description of illustrative
embodiments is intended to be read in connection with the
accompanying drawings, which are to be considered part of the
entire written description. In the description of the exemplary
embodiments disclosed herein, any reference to direction or
orientation is merely intended for convenience of description and
is not intended in any way to limit the scope of the present
invention. The discussion herein describes and illustrates some
possible non-limiting combinations of features that may exist alone
or in other combinations of features. Furthermore, as used herein,
the term "or" is to be interpreted as a logical operator that
results in true whenever one or more of its operands are true.
Furthermore, as used herein, the phrase "based on" is to be
interpreted as meaning "based at least in part on," and therefore
is not limited to an interpretation of "based entirely on."
[0032] As used throughout, ranges are used as shorthand for
describing each and every value that is within the range. Any value
within the range can be selected as the terminus of the range. In
addition, all references cited herein are hereby incorporated by
referenced in their entireties. In the event of a conflict in a
definition in the present disclosure and that of a cited reference,
the present disclosure controls.
[0033] Features of the present invention may be implemented in
software, hardware, firmware, or combinations thereof. The computer
programs described herein are not limited to any particular
embodiment, and may be implemented in an operating system,
application program, foreground or background processes, driver, or
any combination thereof. The computer programs may be executed on a
single computer or server processor or multiple computer or server
processors.
[0034] Processors described herein may be any central processing
unit (CPU), microprocessor, micro-controller, computational, or
programmable device or circuit configured for executing computer
program instructions (e.g., code). Various processors may be
embodied in computer and/or server hardware of any suitable type
(e.g., desktop, laptop, notebook, tablets, cellular phones, etc.)
and may include all the usual ancillary components necessary to
form a functional data processing device including without
limitation a bus, software and data storage such as volatile and
non-volatile memory, input/output devices, graphical user
interfaces (GUIs), removable data storage, and wired and/or
wireless communication interface devices including Wi-Fi,
Bluetooth, LAN, etc.
[0035] Computer-executable instructions or programs (e.g., software
or code) and data described herein may be programmed into and
tangibly embodied in a non-transitory computer-readable medium that
is accessible to and retrievable by a respective processor as
described herein which configures and directs the processor to
perform the desired functions and processes by executing the
instructions encoded in the medium. A device embodying a
programmable processor configured to such non-transitory
computer-executable instructions or programs may be referred to as
a "programmable device", or "device", and multiple programmable
devices in mutual communication may be referred to as a
"programmable system." It should be noted that non-transitory
"computer-readable medium" as described herein may include, without
limitation, any suitable volatile or non-volatile memory including
random access memory (RAM) and various types thereof, read-only
memory (ROM) and various types thereof, USB flash memory, and
magnetic or optical data storage devices (e.g., internal/external
hard disks, floppy discs, magnetic tape CD-ROM, DVD-ROM, optical
disk, ZIP.TM. drive, Blu-ray disk, and others), which may be
written to and/or read by a processor operably connected to the
medium.
[0036] In certain embodiments, the present invention may be
embodied in the form of computer-implemented processes and
apparatuses such as processor-based data processing and
communication systems or computer systems for practicing those
processes. The present invention may also be embodied in the form
of software or computer program code embodied in a non-transitory
computer-readable storage medium, which when loaded into and
executed by the data processing and communications systems or
computer systems, the computer program code segments configure the
processor to create specific logic circuits configured for
implementing the processes.
[0037] Compositions may include one or more ingredients. For
example, a composition may include a first, second, third, etc.
ingredient. One or more of the ingredients of a composition may
have an effect on one or more other ingredients of the composition.
Also, or alternatively, one or more of the ingredients may have an
effect on the composition (e.g., the composition as a whole).
[0038] A composition may be a chemical composition. The chemical
composition may form a product. The chemical composition (e.g.,
product formed from the chemical composition) may be used for one
or more purposes. For example, a product formed from a chemical
composition may be used for cooking; cleaning; personal care;
treating/testing for diseases, disorders, conditions; as well as
one or more other purposes. A composition (e.g., chemical
composition) may be used for performing tasks. For example, a
chemical composition may be used for performing tests, such as
water purity tests.
[0039] A chemical composition may form a personal care product,
although a personal care product is for illustration purposes only
and a chemical composition may form one or more other products. A
personal care product may exist for enhancing a user's health,
hygiene, appearance, etc. Such personal care products may comprise
one or more chemical compositions that are comprised of one or more
ingredients. Personal care products may include oral care products
comprising oral care compositions, skin care products comprising
skin care compositions, hair care products comprising hair care
compositions, as well as other products and/or chemical
compositions.
[0040] Oral care composition, as used herein, may include a
composition for which the intended use can include oral care, oral
hygiene, oral appearance, or for which the intended use may
comprise administration to the oral cavity. Skin care composition,
as used herein, may include a composition for which the intended
use may include promotion or improvement of health, cleanliness,
odor, appearance, and/or attractiveness of skin. Hair care
compositions, as used herein, may include a composition for which
the intended use may include promotion or improvement of health,
cleanliness, appearance, and/or attractiveness of hair. The
compositions may be used for a wide variety of purposes, including
for enhancing personal health, hygiene, and appearance, as well as
for preventing or treating a variety of diseases and other
conditions in humans and in animals.
[0041] FIG. 1A shows a table of data associated with an example
composition. The composition may be a chemical composition, such as
chemical composition 100. Chemical composition 100 may form a
product, such as a personal care product. As can be seen from FIG.
1A, chemical composition 100 may include several ingredients. For
example, chemical composition 100 may include glycerin, sodium
lauryl sulfate, zinc citrate, as well as one or more other
ingredients. Each of the ingredients of the chemical composition
100 (e.g., the chemical composition forming a personal care
product) may be included in the personal care product to provide
one or more predefined features. As provided in FIG. 1A, features
of the ingredients may include providing sweetness to the chemical
composition 100, providing stabilizing factors to the chemical
composition 100, etc. For example, sodium lauryl sulfate is an
ingredient of the chemical composition 100 that may be used as a
solubilizing or cleansing agent for the chemical composition
100.
[0042] FIG. 1B shows a table of additional data associated with an
example composition, such as chemical composition 100. In an
example the chemical composition may form a product, such as a
personal care product. The data provided in FIG. 1B may relate to a
consumer perception of a personal care product. As can be seen from
FIG. 1B, a consumer may perceive a chemical composition according
to several categories and/or features. For example, consumers may
have perceptions about the color, stickiness, wetness, ease of use,
sweetness, etc., of a chemical composition. A consumer may have a
preference for a personal care product based on the perception of
one or more features of the chemical composition forming the
personal care product. For example, a consumer may prefer that a
toothpaste have a certain color, that a shampoo have a certain
smell, that a deodorant have a certain dispersibility, etc. One or
more consumers may rate a personal care product based on one or
more perception values that the consumer has regarding the personal
care product. The values of consumer perceptions may be obtained in
a variety of ways, including a survey (e.g., a paper or online
survey), a clinical trial, commercial success of the personal care
product, etc.
[0043] Consumer perceptions of a chemical composition (e.g., a
chemical composition forming a personal care product) may be based
on one or more of the ingredients of the personal care product.
Said another way, a consumer perception value may be affected by
one or more ingredients of the personal care product. For example,
a particular ingredient may cause a personal care product to be
more white or less white, to be more sticky or less sticky, to
cause more of a burning sensation or less of a burning sensation,
etc.
[0044] FIG. 2A shows a depiction of data associated with a
composition (e.g., a chemical composition). The composition may
form a product, such as a personal care product 200. For example,
data associated with personal care product 200 may include an
identity, such as a name of the personal care product, ingredients
of the personal care product, chemoinformatic values of ingredients
of personal care product, or another identifier used to identify
the personal care product. For example, personal care product 200
may be associated with a unique number 212a that may be referenced
by a user and/or computer when referring to the personal care
product. Each personal care product 200 may have one or more other
values and/or properties, such as property 212b. Property 212b may
be a chemical property associated with the chemical composition of
the personal care product. For example, property 212b may be a
physiochemical property of the chemical composition of the personal
care product. The physiochemical property of the chemical
composition may relate to a physical property or a chemical
property of the chemical composition of the personal care product.
For example, property 212b may be a pH value of the personal care
product. The value of property 212b may be affected by one or more
ingredients of the personal care product.
[0045] Property 212b may be a consumer perception of the personal
care product, such as the perceptions shown in FIG. 1B. For
example, consumers may have perceptions about the color,
stickiness, wetness, ease of use, sweetness, etc., of the personal
care product. The consumer perceptions of the personal care product
may be based on one or more of the ingredients of the personal care
product. For example, an ingredient may cause a personal care
product to be more white or less white, to be more sticky or less
sticky, to cause more of a burning sensation or less of a burning
sensation, etc.
[0046] As described herein, a chemical composition may form a
personal care product. The chemical composition may be comprised of
one or more ingredients (e.g., ingredient data), such as
ingredients 222, 232. Each ingredient may include an identity, such
as a name of the ingredient or other identifier used to identify
the ingredient. For example, ingredient 222 may include a name 222b
and/or an identifier 222a. Ingredient 222 may include other
information, such as the percentage of the personal care product
that comprises the ingredient. For example, as shown in 222c,
ingredient 222 (e.g., sodium lauryl sulfate) may be 1.4999% of the
personal care product 200. Ingredient data may include one or more
other properties and/or values of the properties.
[0047] Each ingredient may further be made up of one or more
substances. As shown in FIG. 2A, ingredient 222 may be comprised of
four substances: water 242, sodium sulfate 252, sodium chloride
262, and sodium C12-16 Alkyl Sulfate 272. Data may be associated
with one or more (e.g., each) of the substances. For example, each
substance may include an identity, such as a name 242a, 252a, 262a,
272a of the substance or other identifier used to identify the
substance. As an example, substance 242 may have a name 242a of
water. The substance (such as substance 242, 252, 262, 272) may
include one or more other values, such as the percentage 242b that
the substance makes up of the ingredient, chemoinformatic
properties of the ingredients (e.g., substances of the
ingredients), etc. For example, substance 242 (i.e., water) may
comprise 70% of the ingredient 222 sodium lauryl sulfate, as shown
on 242b. Substance 272 may include chemoinformatic properties such
as a chemical class, an HLB value, a surface area (e.g.,
topological polar surface area), etc. As shown on FIG. 2A, example
information may include a chemical class of Alkyl Sulfate, an HLB
value of 40, and/or a topological polar surface area of 74.8
squared Angstroms. In other examples, however, one or more
substances may have one or more (e.g., different) chemoinformatic
properties having one or more different values.
[0048] As described herein, property 212b of personal care product
may be affected by interactions of one or more of the ingredients
of the chemical composition that is the personal care product 200.
Examples of property 212b may relate to a pH, fluoride (e.g.,
fluoride stability), viscosity (e.g., viscosity stability),
viscoelasticity, abrasion (e.g., stain removal and dentin
abrasion), color, turbidity, analyte concentration, specific
gravity, consumer perception (e.g., sweetness, stickiness,
fragrance), etc., of a personal care product.
[0049] The value of property 212b of the chemical composition of
the personal care product may be determined by experimentally
measuring the value of the property. By experimentally measuring
the property of the personal care product, the actual value of the
property may be determined. The value of the property 212b may be
determined via a mathematical (e.g., thermodynamic) calculation of
the value of the property. For example, a database of personal care
product compositions may be compiled. The compositions may include
one or more compositions. A catalogue (e.g., a hand-evaluated
catalogue) may contain one or more constants (e.g., metal binding
constants, surface acidity constant, etc.), and/or one or more
solubility products, for example. Speciation calculations may be
performed on personal care product compositions. The speciation
calculations may be used to determine the activity of one or more
(e.g., each) ion of a personal care product composition. The
negative log of the hydrogen ion may correspond (e.g., activity
correspond) to a calculated value (e.g., the calculated pH value)
of the personal care product composition.
[0050] The value of the property 212b may be determined by
receiving consumer perceptions of one or more attributes of the
personal care product. For example, clinical consumer trials may be
used to determine consumer perceptions about the personal care
product. The clinical consumer trials may determine how consumers
(e.g., potential consumers) perceive the personal care product. For
example, clinical consumer trials may determine how consumers feel
about the color (e.g., whiteness), stickiness, wetness, sweetness,
fragrance, bitterness, ease of use, etc. of the personal care
product. Consumer perceptions may be determined via other methods,
including surveys (e.g., online and paper surveys), commercial
success, etc. FIG. 1B provides a list of example properties of
personal care product in which consumers may have a perception.
[0051] FIG. 2B is an example table of data relating to a
composition. The composition may be chemical composition that may
form a personal care product. As described herein, examples of
personal care products may include oral care products (e.g., a
toothpaste, mouthwash, etc.), hair products (e.g., a shampoo, hair
gel, etc.), skin products (e.g., moisturizers, soaps, etc.), etc. A
personal care product may include ingredients, such as the example
ingredients named under column 282. For example, the chemical
composition forming the personal care product may include
ingredients such as sorbitol, water, glycol, etc.
[0052] The ingredients (e.g., each of the ingredients) may be
identified in one or more ways. For example, the ingredients may be
identified by name. Also, or alternatively, the ingredients may be
identified by an identification number (e.g., a unique
identification number), such as by the identification numbers
provided under column 280. The identification number may be used by
a user and/or one or more software applications to identify the
ingredient. The identification number may be used to conceal the
true identity of the ingredients, for example, in instances when
the identification of the ingredients is confidential. The
identification numbers may be randomly generated, may be generated
and/or listed in an order (such as an increasing or decreasing
order), etc. Although the table of FIG. 2B shows the
identifications under column 280 as alphanumeric characters, it
will be understood by those of skill in the art that the
identifications may be represented as any combination of numbers,
letters, special characters, etc.
[0053] FIG. 2B further provides values, such as the percentage
values shown in column 284. The percentage values may relate to the
percentage in which the ingredient comprises the chemical
composition forming the personal care product. For example, as
shown in FIG. 2B the ingredient demineralized water comprises
18.296 percent of the chemical composition, Sodium Lauryl Sulfate
powder comprises 1.5 percent of the chemical composition, and
sodium saccharin USP or EP comprises 0.3 percent of the chemical
composition.
[0054] As described herein, personal care products may be formed of
(e.g., formulated using) one or more chemical compositions
comprising one or more ingredients. Formulating personal care
products using more than one chemical composition and/or one or
more ingredients may present a number of challenges. For example,
combining chemical compositions may cause values of properties of
the chemical composition forming the personal care product to
change. As an example, combining two or more ingredients in a
chemical composition may cause the pH value to change. The pH value
may be changed in an unpredictable way, for example, based on the
interaction of the two or more ingredients.
[0055] As adding, removing, and/or mixing ingredients within a
chemical composition may affect values of properties of the
chemical composition, it may be difficult to create personal care
products in which the addition, reduction, or mixing of ingredients
is required. For example, personal care products may be required to
be pharmaceutically and/or cosmetically acceptable for their
intended uses and/or purposes. The intended uses and/or purposes
may be based on a value of a property (e.g., pH) of the chemical
composition. By combining new ingredients to a chemical
composition, or removing ingredients from a chemical composition, a
value of the property (e.g., a value of pH) of the chemical
composition may change such that the chemical composition forming
the personal care product is no longer suitable for the personal
care products intended purposes.
[0056] Chemical compositions forming personal care products may
contain therapeutic active materials that may (e.g., may only)
deliver desired results if the compositions have not exhibited a
chemical degradation. By combining new ingredients to a chemical
composition, or removing ingredients from a chemical composition, a
value of the property of the chemical composition may change such
that the chemical composition forming the personal care product
incurs a chemical degradation. Such a chemical degradation may
cause the personal care product to no longer be suitable for
consumer use.
[0057] Chemical compositions forming personal care products may
contain cosmetically functional materials that may (e.g., may only)
deliver the material to the oral cavity, skin, and/or hair, etc. at
effective levels under the conditions that they are typically used
by the consumer. By combining new ingredients to a chemical
composition, or removing ingredients from a chemical composition, a
value of the property of the chemical composition may change such
that the chemical composition forming the personal care product no
longer performs at the required effective levels.
[0058] Chemical compositions forming personal care products may
(e.g., may only) exhibit an aesthetic appearance for a time period.
Such aesthetic appeal of chemical compositions may be important,
for example, as such aesthetic appeal may have significant effects
on consumer acceptance and usage. By combining new ingredients to a
chemical composition, or removing ingredients from a chemical
composition, a value of the property of the chemical composition
may change such that the chemical composition forming the personal
care product is no longer aesthetically pleasing.
[0059] Chemical compositions forming personal care products may
exhibit one or more attributes that are perceived by a consumer.
For example, chemical compositions forming a personal care product
may exhibit a flavor, sweetness, ease of use, etc., as perceived by
a consumer. By combining new ingredients to a chemical composition,
or removing ingredients from a chemical composition, a value of the
property of the chemical composition may change such that the
chemical composition forming the personal care product affects the
consumer perception of the personal care product. For example, the
value of the property of the chemical composition may be affected
such that the personal care product exhibits a more minty flavor, a
more salty taste, etc.
[0060] As described herein, it may be possible to determine a value
of a property of a chemical composition forming a personal care
product. For example, a value of a property of a personal care
product composition may be experimentally measured, mathematically
calculated, and/or received via clinical consumer trials. However,
such techniques may be time consuming, nontrivial, and/or
impossible, as the personal care product composition may include
dozens (or more) of ingredients. Machine learning techniques may be
used to determine one or more values of properties of a chemical
composition.
[0061] FIG. 3A shows an example process 300 for using machine
learning techniques to determine (e.g., predict) an attribute. The
attribute may include an identity of a composition, ingredients of
a composition, a value of a property of a composition, etc. For
example, the attribute may include an identity of a chemical
composition forming a personal care product, ingredients of a
chemical composition forming a personal care product, a value of a
property (e.g., a pH value, a fluoride stability value, a viscosity
value, an abrasion value, a specific gravity value, a consumer
perception value) of a chemical composition forming a personal care
product, etc. Although the disclosure may describe the
determination (e.g., prediction) of an identity of a chemical
composition forming a product (e.g., personal care product) or a
value of a property of the chemical composition forming the product
(e.g., personal care product), it should be understood that machine
learning techniques may also, or alternatively, be used to
determine (e.g., predict) other values, such as other identities of
products, values of chemoinformatic properties of ingredients of
products, etc.
[0062] At 302, one or more identities of chemical compositions
(e.g., sample chemical compositions of a personal care product) may
be stored, for example, in a database. The identity of the chemical
composition may include a name of a chemical composition,
ingredients of the chemical composition (e.g., formulations of the
chemical composition), etc. For example, as shown in FIG. 2A, the
identity of the chemical composition may include chemoinformatic
properties (e.g., chemoinformatic values) of each of the
ingredients of the chemical composition. The identity of a chemical
composition may be received from one or more of the databases.
[0063] At 303, one or more perceptions of chemical compositions
forming one or more products (e.g., personal care products) may be
determined and/or received. The perceptions of the chemical
compositions may be determined and/or identified via consumers
(e.g., potential consumers). The perceptions may be determined
and/or identified via clinical consumer trials, for example. The
perceptions may include the whiteness of the personal care product,
how minty the personal care product is, the sweetness of the
personal care product, etc. The perceptions of the chemical
compositions may be affected by one or more ingredients of the
chemical composition forming the personal care product. For
example, one or more ingredients may affect how minty a consumer
perceives the personal care product to be, how sweet the consumer
perceives the personal are product to be, how white the consumer
perceives the personal are product to be, etc.
[0064] At 304, values of properties of chemical compositions
forming one or more products (e.g., personal care products) may be
determined via experimental measurements. The values of the
properties may be affected by one or more ingredients of the
chemical composition. The experimentally measured values of
properties of the chemical composition may be identified by
performing actual measurements of the values of the properties of
the chemical compositions. The experimentally measured values of
properties of the chemical composition may be identified by
retrieving the experimentally measured values of the properties
from a database, for example, after the experimentally measured
values of the properties have been stored in a database. The
experimentally measured values of properties of the chemical
composition (e.g., sample chemical compositions) may be
received.
[0065] At 306, one or more values of a chemical composition forming
a personal care product may be determined and/or stored, for
example, in a database. The one or more values of the chemical
composition may relate to physiochemical properties of a chemical
composition. The value of a physiochemical property may include a
value for one or more (e.g., each) ingredients of the chemical
composition. The value of a physiochemical property may be
received, for example, from one or more databases.
[0066] At 308, the values of physiochemical properties of the
chemical composition may be identified and/or determined. The
values of physiochemical properties of the chemical composition may
be determined by measuring the physiochemical properties of the
ingredients of the chemical composition, calculating (e.g.,
mathematically calculating) predicted values of the physiochemical
properties of the chemical compositions, looking up the values of
the physiochemical properties (e.g., via a database, look-up table,
etc.), etc. The values of physiochemical properties of the chemical
composition may be identified and/or determined via thermodynamic
calculations of the physiochemical properties.
[0067] At 310, data may be input into a machine learning model, as
described herein. For example, identities of chemical compositions
may be input into the model. Identities of chemical compositions
may include names of one or more of the chemical compositions,
identities of ingredients of the chemical compositions, values of
chemoinformatic properties (of ingredients) of the chemical
compositions, etc. Values of properties of the chemical composition
may be input into the machine learning model. For example, values
of properties (e.g., experimentally measured values, mathematically
calculated values, consumer perceived values) of the chemical
composition may be input into the model. Data related to chemical
compositions may be input into the model to train the model, in
examples. In other examples, data related to chemical compositions
may be input into the model to determine values (e.g., other
values) of the chemical compositions.
[0068] An association may be input into the model. For example,
there may be an association between an identity (e.g., ingredients)
of a chemical composition and a value of a property of the chemical
composition. The ingredients of a chemical composition (e.g., a
sample chemical composition) and the associated value of the
property of the chemical composition (e.g., a sample chemical
composition) may be input into a machine learning model, for
example, to train the machine learning model.
[0069] At 312, the machine learning model may determine (e.g.,
predict) a value of one or more pieces of data relating to the
chemical composition. For example, if an identity of a chemical
composition (e.g., a considered chemical composition) is input into
the machine learning model, the machine learning model may
determine (e.g., predict) a value of a property of the chemical
composition based on the identity of the chemical composition.
Conversely, if a value of a property of a chemical composition
(e.g., considered chemical composition) is input into the machine
learning model, the machine learning model may determine (e.g.,
predict) an identity of the chemical composition based on the value
of the property of the chemical composition.
[0070] FIG. 3B is an example diagram of system 350 for determining
information relating to a composition, such as a chemical
composition forming a product (e.g., a personal care product). The
information may relate to an identity of a chemical composition,
chemoinformatic values of the chemical composition, as well as one
or more other values relating to a product. System 350 may be a
data warehouse, in an example. For example, system 350 may include
one or more databases for receiving, storing, and/or providing data
and/or one or more processors for processing the data received,
stored, and/or provided by one or more of the databases.
[0071] System 350 may include element 352, which may include one or
more databases. For example, element 352 may include one or more
databases receiving, storing, and/or providing formulation
identifiers, raw materials in one or more (e.g., each) of the
formulations, and/or weight percentages of one or more (e.g., each)
of the raw materials in a formulation. Element 352 may include one
or more databases receiving, storing, and/or providing formulation
identifiers, descriptive sales, and/or logistical information.
Element 352 may include one or more databases receiving, storing,
and/or providing raw material identifiers, cost(s), manufacturer
information, and/or logistical information. Element 352 may include
one or more databases receiving, storing, and/or providing raw
material identifiers, chemicals in one or more (e.g., each) raw
material, and/or weight percentages of one or more (e.g., each)
chemical in a raw material. Element 352 may include one or more
databases receiving, storing, and/or providing raw material
identifiers and/or informatic (e.g., chemoinformatic) properties of
the raw materials. Element 352 may include one or more databases
receiving, storing, and/or providing chemical identifiers and/or
informatic (e.g., chemoinformatic) properties of the chemicals.
Element 352 may include one or more databases receiving, storing,
and/or providing thermodynamic and kinetic reaction constants
between chemicals, such as all known thermodynamic and kinetic
reaction constants between all chemicals.
[0072] At 354, feature selection, representation, and/or
engineering may be performed. For example, rules (e.g., algorithms)
may perform feature selection, representation, and/or
engineering.
[0073] System 350 may include element 356, which may include one or
more databases. For example, element 356 may include one or more
databases receiving, storing, and/or providing formulation
identifiers, select features in a (e.g., each) formulation (such as
a combination of identifiers, material informatics, chemical
informatics, etc.), and/or representation (e.g., quantitative
representation) of the abundance of a feature in a formulation.
[0074] At 362, chemical speciation calculations (e.g., based on
thermodynamic and/or kinetic constants) may be performed. For
example, rules (e.g., algorithms) may perform chemical speciation
calculations (e.g., based on thermodynamic and/or kinetic
constants).
[0075] System 350 may include element 364, which may include one or
more databases. For example, element 364 may include one or more
databases receiving, storing, and/or providing formulation
identifiers, calculated values of a property of a chemical
composition, and/or equilibrium properties (e.g., based on kinetics
and thermodynamic constants).
[0076] System 350 may include element 366, which may include one or
more databases. For example, element 366 may include one or more
databases receiving, storing, and/or providing formulation
identifiers and/or testing values (e.g., experimentally determined
analytical testing values of a property of a chemical composition).
The property of the sample chemical composition may be affected by
an interaction two or more of the ingredients of the sample
chemical composition. Element 366 may include one or more databases
receiving, storing, and/or providing formulation identifiers and/or
consumer-derived testing results. Element 366 may include one or
more databases receiving, storing, and/or providing formulation
identifiers and/or clinical testing results.
[0077] At 368, fitting parameters of testing results may be
determined. For example, rules (e.g., algorithms) may determine
fitting parameters of testing results.
[0078] System 350 may include element 370, which may include one or
more databases. For example, element 370 may include one or more
databases receiving, storing, and/or providing formulation
identifiers, aggregated testing results, and/or fitting parameters
associated with testing results.
[0079] At 358, machine learning information may be determined. For
example, rules (e.g., algorithms) may determine machine learning
information.
[0080] System 350 may include element 360, which may include one or
more databases. For example, element 360 may include one or more
databases receiving, storing, and/or providing machine learning
model parameters.
[0081] FIG. 4 is a process 400 showing other example steps of
predicting chemical composition information via machine learning
rules, as described herein. Hatch lines are used in FIG. 4 to
denote relationships within the process.
[0082] At 402, the process begins. At 404, an entity (e.g., a
business) may begin to understand and/or improve its understanding
of chemical composition information. For example, the entity may
begin and/or improve its understanding of a need for chemical
compositions to have a feature, such as a chemical composition
having a pH of a certain value, an emulsifying purpose, a
sweetness, a thickener, etc. Although the entity may understand a
need for the chemical composition to have a certain feature (e.g.,
value), the entity may not know the ingredients of the chemical
composition that will create such a feature (e.g., value).
[0083] At 406, data relating to the chemical composition may be
acquired. For example, the entity may acquire identities (e.g.,
names, ingredients, chemoinformatic properties, etc.) of chemical
compositions, values of properties of chemical compositions, etc.
The information of the chemical composition may be acquired via an
experimental measurement, a mathematical computation, clinical
consumer trials, one or more data sources (e.g., a database, file,
etc.), or other informational avenues. An association between the
information may be identified and/or determined. For example, an
association between an identity of a chemical composition and a
value of a property of the chemical composition may be
determined.
[0084] At 408, a machine learning model may be trained and/or used,
as described herein. For example, information relating to a
chemical composition (e.g., a sample chemical composition) may be
used to train a machine learning model. The information may be an
identity of a chemical composition and an associated value of a
property of the chemical composition. The trained machine learning
model may be used to determine and/or predict a value (e.g., an
unknown value) of a chemical composition (such as a considered
chemical composition), for example, based on an identity of the
chemical composition (e.g., the chemical composition).
[0085] At 410, the machine learning model may be deployed. At
deployment, the machine learning model may determine a value of a
property of a chemical composition (e.g., a considered chemical
composition) based on an identity of the chemical composition, may
determine an identity of the chemical composition based on a value
of a property of a chemical composition, etc.
[0086] The determined value of the property of the chemical
composition may be compared against a desired value of the property
of the chemical composition. For example, the pH value returned
from the machine learning model may be compared against a desired
pH value. The pH value returned from the machine learning model may
be compared against an actual (e.g., actually measured) pH value.
If it is determined that the value of the property is the same
(e.g., substantially the same) as the desired value, the entity may
move towards creating a chemical composition (e.g., a personal care
product) having the desired value of the property. The entity may
use the ingredients input into the machine learning model to create
the chemical composition having the desired value of the property.
For example, the entity may create a chemical composition using the
ingredients input into the machine learning model that resulted in
a determined (e.g., predicted) pH value that is desired. A personal
care product may be created using the chemical composition such
that the personal care product will be comprised of ingredients
resulting in the desired value of the property.
[0087] FIG. 5 is a block diagram of an example system 500 for
determining (e.g., predicting) data associated with a composition
(e.g., chemical composition) forming a product, such as a personal
care product. Data may include an identity of a chemical
composition, a value of a property of a chemical composition,
and/or one or more types of data. The data may be determined based
on one or more attributes and/or parameters. For example, system
500 may determine (e.g., predict) the data associated with a
property of a chemical composition based on one or more
attributes/parameters and machine learning techniques. Although
examples provided herein may relate to determining (e.g.,
predicting) an identity of a chemical composition, a value of a
property of a chemical composition, and/or a fitting parameter
using machine learning techniques, a person of skill in the art
will understand that one or more other values and/or parameters
relating to a chemical composition may be determined (e.g.,
predicted) using machine learning techniques. For example,
chemoinformatic values of ingredients of a chemical composition may
be determined, chemical constants may be determined, consumer
perceptions may be determined, etc.
[0088] System 500 includes a user device 502 configured to connect
to a properties modeling device, such as example chemical
properties modeling device 602 (further described in FIG. 6) via a
network 520. Network 520 may include wired and/or wireless
communication networks. For example, networks 520 may include a
local area network (LAN), a metropolitan area network (MAN), and/or
a wide area network (WAN). Network 520 may facilitate a connection
to the Internet. In further examples, network 520 may include wired
telephone and cable hardware, satellite, cellular phone
communication networks, etc.
[0089] User device 502 may include a user interface 504, a memory
506, a central processing unit (CPU) 508, a graphics processing
unit (GPU) 510, an image capturing device 514, and/or a display
512. User device 502 may be implemented as a user equipment (UE)
such as a mobile device, a computer, laptop, tablet, desktop, or
any other suitable type of computing device.
[0090] User interface 504 may allow a user to interact with user
device 502. For example, user interface 504 may include a
user-input device such as an interactive portion of display 512
(e.g., a "soft" keyboard displayed on display 512), an external
hardware keyboard configured to communicate with user device 504
via a wired or a wireless connection (e.g., a Bluetooth keyboard),
an external mouse, or any other user-input device. The user
interface 504 may allow a user to input, view, etc. one or more
pieces of information relating to a chemical composition forming a
personal care product.
[0091] Memory 506 may store instructions executable on the CPU 508
and/or the GPU 510. The instructions may include machine readable
instructions that, when executed by CPU 508 and/or GPU 510, cause
the CPU 508 and/or GPU 510 to perform various acts. Memory 506 may
store instructions that when executed by CPU 508 and/or GPU 510
cause CPU 508 and/or GPU 510 to enable user interface 504 to
interact with a user. For example, executable instructions may
enable user interface to display (via Display 512) one or more
prompts to a user, and/or accept user input. Instructions stored in
memory 506 may enable a user to input an identity of a chemical
composition and/or a value of a property of the chemical
composition, for example. In other examples, a user may utilize
user interface 504 to click, hold, or drag a cursor to define
identities, values, and/or properties of a chemical
composition.
[0092] CPU 508 and/or GPU 510 may be configured to communicate with
memory 506 to store to and read data from memory 506. For example,
memory 506 may be a computer-readable non-transitory storage device
that may include any combination of volatile (e.g., random access
memory (RAM)) or non-volatile (e.g., battery-backed RAM, FLASH,
etc.) memory.
[0093] Image capturing device 514 may be configured to capture an
image. The image may be a two-dimensional image, a
three-dimensional image, etc. Image capturing device 514 may be
configured to capture an image in a digital format having a number
of pixels. Although image capturing device 514 is illustrated in
FIG. 5 as internal to user device 502, in other examples image
capturing device 514 may be internal and/or external to user device
502. In an example, image capturing device 514 may be implemented
as a camera coupled to user device 502. Image capturing device 514
may be implemented as a webcam coupled to user device 502 and
configured to communicate with user device 502. Image capturing
device 514 may be implemented as a digital camera configured to
transfer digital images to user device 502 and/or to chemical
properties modeling device 602. Such transfers may occur via a
cable, a wireless transmission, network 520/620, and/or a physical
memory card device transfer (e.g., SD Card, Flash card, etc.), for
example. Image capturing device 514 may be used to capture an image
of a personal care product, a chemical composition forming the
personal care product, data relating to the chemical composition,
data relating to one or more features of a personal care product,
etc.
[0094] In examples the user may input information into the user
device 502 relating to one or more compositions (e.g., chemical
compositions). The chemical composition information may be
transferred to and/or from the chemical property modeling device
602, as shown in FIG. 5. With the chemical property modeling device
602 having information relating to the chemical compositions (e.g.,
the identities of the chemical compositions and/or the values of
the properties of the chemical compositions), the chemical property
modeling device 602 may return information about the chemical
composition. For example, the chemical property modeling device 602
may provide values (e.g., predicted values) of properties of
chemical compositions.
[0095] User device 502 may obtain information (e.g., unknown
information) about one or more chemical compositions (e.g., names
of chemical compositions, ingredients of chemical compositions,
chemoinformatic values of ingredients of chemical compositions,
values of properties of chemical compositions, etc.) for prediction
purposes. For example, a user (e.g., a user of user device 502) may
desire to know an identity of a chemical composition having a value
(e.g., desired value) of a property of a personal care product. The
value of the property may be affected by one or more ingredients of
the chemical composition interacting within one another. The value
(e.g., desired value) of a property of a personal care product may
be a value (e.g., a pH value) of the personal care product, a
function of one or more ingredients of the personal care product, a
classification of one or more ingredients of the personal care
product, a consumer perception of the personal care product,
etc.
[0096] The user may input one or more types and/or values of
chemical composition information (e.g., names, ingredients,
chemoinformatic properties, etc.) into the user device 502, for
example, to determine information (e.g., other information) about
the chemical compositions. The user device 502 may transmit the
information to chemical property modeling device 602. In examples
all or some of the steps, processes, methods, etc., may be
performed by one device or more than one device (e.g., user device
or chemical property modeling device). For example, user device 502
may include chemical properties engine 630 in examples. In other
examples, chemical property modeling device 602 may be external to
user device 502.
[0097] In examples in which chemical property modeling device 602
is separate from user device 502, user device 502 may communicate
with chemical property modeling device 602 via one or more wired
and/or wireless techniques, as described herein. For example, as
shown in FIG. 5, user device 502 may communicate with chemical
property modeling device 502 via network 520. Network 520 may be
the Internet, in some examples. In other examples, as described
herein, network 520 may be Wi-Fi, Bluetooth, LAN, etc.
[0098] A value of a property (e.g., a desired value of a property)
of a chemical composition may be received. The chemical composition
in which the value of the property is received and in which the
identity of the chemical composition is to be determined by machine
learning rules may be referred to as a considered chemical
composition. For example, a user may receive a value (e.g., a
desired value) of a property of a chemical composition. The user
may transfer the value of the property to the chemical property
modeling device 602. The value may relate to a property that is
affected by one or more ingredients of the chemical composition,
such as a pH value, a fluoride stability value, a viscosity value,
an abrasion value, a specific gravity value, a consumer perception
value, etc.
[0099] Based on the value of the property, the chemical property
modeling device 602 may provide an identity of a chemical
composition that has (e.g., is predicted to have) that value (e.g.,
or approximately that value) for the property. For example, the
chemical property modeling device 602 may provide a name of a
composition that has (e.g., is predicted to have) the value,
ingredients of a chemical composition that has (e.g., is predicted
to have) the value, chemoinformatic values of ingredients of a
chemical composition that has (e.g., is predicted to have) the
value, etc.
[0100] The user may also, or alternatively, provide information
related to the chemical composition to determine a value of a
property of the chemical composition. For example, the user may
input into the chemical property modeling device 602 a name of a
composition, ingredients of a chemical composition, chemoinformatic
values of ingredients of a chemical composition, consumer
perceptions of the chemical composition, etc. Based on the name,
ingredient, and/or chemoinformatic information, the chemical
property modeling device 602 may determine a value of a property of
the chemical composition. For example, based on the name,
ingredient, and/or chemoinformatic information, the chemical
property modeling device 602 may determine a pH value, a fluoride
stability value, a viscosity value, an abrasion value, a specific
gravity value, etc., of the chemical composition.
[0101] FIG. 6 shows an example system 600 of training a properties
engine, such as chemical properties engine 630. Chemical properties
engine 630 may be housed in chemical property modeling device 602,
although such a configuration is for illustration purposes only. As
shown in FIG. 6, training device 650 may communicate with chemical
property modeling device 602. For example, training device 650 may
communicate with chemical property modeling device 602 via network
620. One or more training devices 650 may provide information to
the chemical property modeling device 602, for example, to train
the chemical properties engine 630 of chemical property modeling
device 602, as described herein.
[0102] Training device 650 may provide information to a modeling
device, such as chemical property modeling device 602. Information
provided to chemical property modeling device 602 may include
experimentally measured information relating to a chemical
composition (e.g., a chemical composition forming a personal care
product), mathematically calculated information relating to a
chemical composition, consumer perception information relating to a
chemical composition, etc. Training device 650 may provide
information relating to chemical compositions that includes
identities of chemical compositions (e.g., names of chemical
compositions, ingredients of chemical compositions, chemoinformatic
values of ingredients of chemical compositions, etc.). The training
device 650 may provide values of properties of chemical
compositions, such as actual values of properties of chemical
compositions and/or mathematically determined values of properties
of chemical compositions.
[0103] As provided herein, information provided by training device
650 may be based on actual (e.g., actually measured information,
such as values of chemical compositions having actually been
measured). In addition, or alternatively, information provided by
training device 650 may be based on values of the chemical
compositions being determined using mathematical calculations, such
as thermodynamic calculations of the chemical compositions to
determine values of properties of the chemical compositions.
Providing this information (e.g., actual information and/or
thermodynamically calculated information) to the chemical
properties engine 630 may be used to train the model, using machine
learning techniques, as described herein. The chemical composition
for which information is used to train the machine learning rules
may be referred to as a sample chemical composition.
[0104] Chemical property modeling device 602 may include a CPU 608,
memory 606, GPU 610, interface 616, and chemical properties engine
630. Memory 606 may be configured to store instructions executable
on the CPU 608 and/or the GPU 610. The instructions may include
machine readable instructions that, when executed by CPU 608 and/or
GPU 610, cause the CPU 608 and/or GPU 610 to perform various acts.
CPU 608 and/or GPU 610 may be configured to communicate with memory
606 to store to and read data from memory 606. For example, memory
606 may be a computer-readable non-transitory storage device that
may include any combination of volatile (e.g., random access memory
(RAM), or a non-volatile memory (e.g., battery-backed RAM, FLASH,
etc.) memory.
[0105] Interface 616 may be configured to interface with one or
more devices internal or external to chemical property modeling
device 602. For example, interface 616 may be configured to
interface with training device 650 and/or chemical properties
database 624. Chemical properties database 624 may store
information about chemical compositions, such as names of chemical
compositions, ingredients of chemical compositions, chemoinformatic
values of ingredients of chemical compositions, values of
properties of chemical compositions (e.g., pH values, fluoride
(e.g., fluoride stability) values, viscosity (e.g., viscosity
stability) values, abrasion (e.g., stain removal and dentin
abrasion) values, specific gravity values, consumer perception
(e.g., sweetness, stickiness, fragrance) values), etc. The
information stored within chemical properties database 624 may be
used to train the chemical properties engine 630. The information
stored within chemical properties database 624 may also, or
alternatively, be referenced by chemical properties engine 630 for
determining (e.g., predicting) information about a chemical
composition (e.g., a considered chemical composition).
[0106] A device (e.g., user device 502 and/or chemical property
modeling device 602) may receive information of one or more
chemical compositions via training device 650 and/or another
device. The information may relate to one or more (e.g., many)
different types of chemical compositions, a family of chemical
compositions, complete and/or incomplete chemical compositions,
chemical compositions with extensive history, relatively unknown
chemical compositions, etc.
[0107] One or more types of information of a chemical composition
may be provided to chemical property modeling device 602. For
example, one or more types of information of a chemical composition
(e.g., sample chemical composition) may be provided to chemical
property modeling device 602 to train the chemical property
modeling device 602 (e.g., machine learning rules of the chemical
property modeling device 602). For example, for a (e.g., each)
chemical composition, the chemical property modeling device 602 may
receive actual (e.g., actually measured) information of the
chemical composition, calculated (e.g., thermodynamically
calculated) information of the composition, predicted information
of the chemical composition, identity information of the chemical
composition, consumer preference information of the chemical
composition, etc. The chemical property modeling device 602 may
perform an association of the information so that a prediction of
chemical composition data (e.g., similar chemical composition data)
may be performed.
[0108] Chemical property modeling device 602 may use machine
learning techniques to develop a software application (e.g., a
model). For example, chemical properties engine 630 may include
machine learning rules for determining (e.g., predicting)
information relating to a chemical composition. Chemical properties
engine 630 may include a model (e.g., a machine learning model) to
determine (e.g., predict) information regarding a chemical
composition. The information provided to the model and/or the
information provided by the model may be used to train the model.
The information used to train the model may include identities
(e.g., names, ingredients, chemoinformatic values of ingredients,
etc.) of a chemical composition, values of properties of the
chemical composition, consumer perception information of the
chemical composition, etc. The information provided to and/or by
the model to train the model may relate to chemical compositions
(e.g., sample chemical compositions).
[0109] The chemical properties engine 630 may include currently
known and/or later developed machine learning rules or algorithms.
The machine learning rules may be supervised machine learning rules
and/or unsupervised machine learning rules. For example, the
chemical properties engine 630 may include at least one of a Random
Forest rule, Support Vector Machine rule, Naive Bayes
Classification rule, Boosting rule, a variant of a Boosting rule,
Alternating Decision Tree rules, Support Vector Machine rules,
Perceptron rules, Winnow rules, Hedge rules, rules constructing a
linear combination of features or data points, Decision Tree rules,
Neural Network rules, logistic regression rules, log linear model
rules, Perceptron-like rules, Gaussian process rules, Bayesian
techniques, probabilistic modeling techniques, regression trees,
ranking rules, Kernel Methods, Margin based rules,
linear/quadratic/convex/conic/semi-definite programming techniques,
or any modifications of the foregoing.
[0110] The chemical properties engine 630 may improve its ability
to perform a task as it analyzes more data related to the task. As
described herein, the task may be to determine (e.g., predict)
unknown information relating to a chemical composition forming a
personal care product. The unknown information may be an unknown
value of a property of a chemical composition, for example, from
known information. For example, the task may be to predict the
value of a property of a chemical composition based on identity
information of the chemical composition. Conversely, the task may
be to predict the identity of a chemical composition based on a
value of a property of the chemical composition. In such examples,
the more information (relating to one or more chemical
compositions) provided to the model, the better the results from
the model may be. For example, the model may provide more accurate
determinations of values of properties of chemical compositions
based on the model receiving numerous pieces of information of the
chemical compositions and information related to the identities of
the chemical compositions.
[0111] As described herein, the machine learning model may be
trained using a set of training examples. Each training example may
include an example of an object, along with a value for the
otherwise unknown property of the object. By processing a set of
training examples that include the object and/or the property value
for the object, the model may determine (e.g., learn) the
attributes or characteristics of the object that are associated
with a particular property value. This learning may then be used to
predict the property or to predict a classification for other
objects. As described herein, machine learning techniques (e.g.,
rules, algorithms, etc.) may be used to develop models for one or
more chemical compositions.
[0112] Chemical compositions (and/or one or more ingredients of the
chemical compositions) forming a personal care product may be
identified and/or classified based on product, function,
classification, consumer perception, etc. One or more chemical
compositions (and/or one or more ingredients within the chemical
compositions) may be identified and/or classified prior to the
chemical compositions being input into the machine learning rules.
One or more chemical compositions (and/or one or more ingredients
within the chemical compositions) may be identified and/or
classified by the machine learning rules. For example, machine
learning rules may identify and/or classify chemical compositions
(and/or one or more ingredients within the chemical compositions)
based on product, function, classification, consumer perception,
etc.
[0113] Models (e.g., machine learning models) may be developed to
receive information relating to a chemical composition, for
example, to determine (e.g., predict) information of a chemical
composition. Training examples (e.g., training sets or training
data) may be used to train the chemical properties engine 630. For
example, the training data may include the names of sample chemical
compositions, ingredients of sample chemical compositions,
chemoinformatic values of ingredients of sample chemical
compositions, fitting parameters of sample chemical compositions,
functions of sample chemical compositions, classifications of
sample chemical compositions, values of properties of sample
chemical compositions, etc. The values of properties of sample
chemical compositions may be determined via calculations, such as
via thermodynamic calculations. The values of properties of sample
chemical compositions may be determined via an experimental
measurement. As described herein, properties of sample chemical
compositions may include the pH, fluoride stability, viscosity
stability, abrasion, specific gravity, consumer perception
properties, etc. of the sample chemical composition.
[0114] After training the chemical properties engine 630 (e.g., the
machine learning model of chemical properties engine 630) using
training data, the chemical properties engine 630 may be used to
determine (e.g., predict) data. For example, the chemical
properties engine 630 may be used to determine (e.g., predict)
parameters that are similar to the parameters used to train the
chemical properties engine 630. As an example, chemical properties
engine 630 may be trained using identities (e.g. ingredients) of
chemical compositions and values of a pH property of chemical
compositions. The chemical properties engine 630 may be used to
determine unknown values of a pH property, for example, based on
the identity (e.g., ingredients) of the chemical composition.
[0115] In other examples, after training the chemical properties
engine 630 (e.g., the machine learning model of chemical properties
engine 630) using training data, the chemical properties engine 630
may be used to determine parameters that are different than the
parameters used to train the chemical properties engine 630. As an
example, chemical properties engine 630 may be trained using
identities (e.g. ingredients) of chemical compositions and values
of a pH property of chemical compositions. The chemical properties
engine 630 may be used to determine unknown values of a soluble
zinc property. The chemical properties engine 630 may be used to
determine unknown values of a soluble zinc property based on the
identity (e.g., ingredients) of the chemical composition. The
different parameters may have a relationship with one another. The
relationship between the different parameters may allow the
chemical properties engine 630 to predict the different parameters.
Using the example above, although the pH property and the soluble
zinc property are different properties, there may be a relationship
with the pH property and the soluble zinc property that allows the
chemical properties engine 630 to predict the soluble zinc data
based on pH training data.
[0116] Other data related to a chemical composition may be used to
train the chemical properties engine 630. For example, data related
to a property may be used to train the chemical properties engine
630. As an example, the data related to the property may be a
fitting parameter associated with the value of the property,
although other types of data may be used to train the chemical
properties engine 630. As an example, training data may include an
identity (e.g., a name, ingredients, and/or chemoinformatic values
of ingredients) of sample chemical compositions and fitting
parameters of sample chemical compositions.
[0117] A fitting parameter may be used to determine a value of a
parameter at a defined instance. For example, the fitting parameter
may relate to the rate at which a value changes over time. The
fitting parameter may be used to determine a value of a parameter
at a future date, day, time, time period, etc. A fitting parameter
may be used to define continuous functions. A fitting parameter may
be used to determine a value of a property at one or more (e.g.,
any) point in time. For example, if a value of fluoride stability
has been measured at 4, 8, and 13 weeks, a fitting parameter may be
derived which may provide values (e.g., expected values) of
fluoride stability at intermediate timepoints between 4, 8, and 13
weeks and/or at extended points beyond 13 weeks.
[0118] Determining a value of a property at a future date, day,
time, time period, etc. may be useful as manufacturers of products
(e.g., personal care products) may be required to demonstrate that
a product (e.g., personal care product) maintains a minimum
threshold quantity of a property throughout the shelf-life of the
product. For example, as shelf lives of a product may be on the
order of several years, it may be impractical to test products
(e.g., new products) at certain time periods (e.g., months, years,
etc.) to determine the viability of the products. It may be useful
to collect data (e.g., collect data over a short period of time)
and use a fitting parameter to extrapolate the value of the
property at a longer period of time. Such a model (e.g., a model
which predicts fitting parameters) may predict properties at
timepoints for which there may be no experimental data.
[0119] As an example, a chemical properties engine 630 may be
trained using data related to a chemical composition. As described
herein, the data may be an identity (e.g., ingredients) of the
chemical composition as well as other data. For example, the
chemical properties engine 630 may be trained using ingredients of
chemical compositions, molecular weights of the ingredients (e.g.,
each ingredient), weight percentages of the ingredients (e.g., each
ingredient), etc. The weight percentages of the ingredients (e.g.,
each ingredient) may be converted to a molar concentration. The
chemical properties engine 630 may be trained using molar
concentrations, theoretical total fluoride content, and/or soluble
fluoride after aging of a chemical composition (e.g., after aging
of a chemical composition for 13 weeks at 40 degrees Celsius).
After training the chemical properties engine 630 (e.g., the
machine learning model of chemical properties engine 630) using
training data, the chemical properties engine 630 may be used to
determine (e.g., predict) data. For example, the chemical
properties engine 630 may be used to determine (e.g., predict) a
value for soluble fluoride after aging, based on an identity (e.g.,
ingredients) of a chemical composition and/or based on molecular
concentration data related to the chemical composition.
[0120] A fitting parameter may be used to determine a value of a
parameter at a future time period. For example, the chemical
properties engine 630 may be trained using ingredients of chemical
compositions, molecular weights of the ingredients (e.g., each
ingredient), weight percentages of the ingredients (e.g., each
ingredient), and/or the fitting parameter. After training the
chemical properties engine 630 (e.g., the machine learning model of
chemical properties engine 630) using ingredients of chemical
compositions, molecular weights of the ingredients (e.g., each
ingredient), weight percentages of the ingredients (e.g., each
ingredient), and/or the fitting parameter, the chemical properties
engine 630 may be used to determine the fitting parameter. The
fitting parameter may be used with a fitting function to determine
a defined instance, as described herein. For example, the chemical
properties engine 630 may be used to determine (e.g., predict) a
value of a fitting parameter that can be used with a fitting
function to determine the theoretical total fluoride content and/or
the soluble fluoride content measured after aging of a chemical
composition (e.g., after aging of a chemical composition at 4, 8,
and/or 13 weeks at 40 degrees Celsius). Example fitting functions
may include an exponential function, polynomial function, power
function, trigonometric function, although other fitting functions
may be used.
[0121] Information included in the training data may be selected
and/or input into the model of chemical properties engine 630 based
on a function and/or classification. For example, a chemical
composition and/or an ingredient of a chemical composition may have
a defined function and/or classification, such as an ingredient in
a chemical composition having a binding function, a preserving
function, a whitening function, an alcohol classification, an
ethers classification, etc. The function may relate to how one or
more of the ingredients of the chemical composition are used in
product form. The function may relate to a pellicle cleaning ratio
(PCR) and/or a relative dentin abrasivity (RDA). PCR is a
measurement of stain removal and may represent the cleaning
efficacy of a personal care product, such as a toothpaste. RDA is a
measurement of abrasivity (e.g., pure abrasivity) and may represent
the erosive capability of a personal care product, such as a
toothpaste. Example functions 700 of a chemical composition and/or
an ingredient within the chemical composition may be found in FIG.
7. Example classifications 800 of a chemical composition and/or an
ingredient within the chemical composition may be found in FIG. 8.
Although FIG. 7 and FIG. 8 provide a list of functions and
classifications, respectively, it should be understood by those of
skill in the art that the functions provided in FIG. 7 and the
classifications are provided in FIG. 8 are for example purposes
only and are non-limiting.
[0122] In examples, one or more chemical compositions may (e.g.,
may only) be input into a model if the chemical compositions have a
function (e.g., a desired function). As an example, a set may
consist of eighty chemical compositions. Of the eighty chemical
compositions, fourteen chemical compositions may include an
ingredient that provides a function of whitening. A user may desire
to determine a value of a property of a chemical composition
wherein the chemical composition (e.g., an ingredient of the
chemical composition) may have a whitening function. In such an
example, the model may be trained using (e.g., only using) chemical
compositions (e.g., ingredients of the chemical composition) having
a whitening function, such as the fourteen chemical compositions in
the above example. Also, or alternatively, the model may categorize
(e.g., automatically categorize, dynamically categorize, etc.)
chemical compositions based on the function of the chemical
composition.
[0123] With the model including (e.g., only including) chemical
compositions of a defined function, or the model categorizing
chemical compositions based on a defined function, the chemical
properties engine 630 may provide information of a chemical
composition having (e.g., only having) the function. For example,
the chemical properties engine 630 may determine a value of a
property of a chemical composition having a function (e.g., a
flavoring function, a binding function, etc.) based on an identity
of the chemical composition. Conversely, the chemical properties
engine 630 may determine an identity of a chemical composition
having a function based on a value of a property of the chemical
composition.
[0124] In other examples, one or more chemical compositions may
(e.g., may only) be input into a model if the chemical compositions
have a classification (e.g., a chemical classification). The
classification may relate to molecular properties of an ingredient
of a chemical composition, such as a chemical composition forming a
personal care product. The chemical compositions may (e.g., may
only) be input into a model if the chemical compositions have a
desired classification. Example chemical classifications may
include an alcohols classification, an amino acids classification,
an enzymes classification, a fatty acids classification, a ketones
classification, peptides classification, as well as other
classifications provided in FIG. 8.
[0125] As an example, a set may consist of forty chemical
compositions. Of the forty chemical compositions, ten chemical
compositions may include an ingredient that is classified as an
ether. A user may desire to determine a value of a property of a
chemical composition wherein the chemical composition (e.g., an
ingredient of the chemical composition) may have a classification
of an ether. In such an example, the model may be trained using
(e.g., only using) chemical compositions (e.g., ingredients of the
chemical composition) having a classification of an ether. Also, or
alternatively, the model may categorize (e.g., automatically
categorize, dynamically categorize, etc.) chemical compositions
based on the classifications of the chemical composition.
[0126] With the model including (e.g., only including) chemical
compositions of a defined classification, or the model categorizing
chemical compositions based on a defined classification, the
chemical properties engine 630 may provide information of a
chemical composition having (e.g., only having) the classification.
For example, the chemical properties engine 630 may determine a
value of a property of a chemical composition having a
classification (e.g., an alcohols classification, a fatty acids
classifications, etc.) based on an identity of the chemical
composition. Conversely, the chemical properties engine 630 may
determine an identity of a chemical composition having a
classification based on a value of a property of the chemical
composition.
[0127] As described herein, information relating to one or more
chemical compositions may be input into the model based on a
function, classification, consumer perception, etc., of the
chemical composition and/or ingredient of the chemical composition.
Information (e.g., identities, values of properties, etc.) of
chemical compositions may be identified based on experimentation,
simulation, mathematical computations, analysis, clinical consumer
trials, and/or assumptions regarding the property being modeled.
For example, actual (e.g. actually measured) values of properties
of chemical compositions may be identified and input into the
model.
[0128] A training set (e.g., identities of chemical compositions
and associated values of properties of the chemical compositions)
may be used to train a machine learning model (e.g., chemical
properties engine 630). The machine learning model (e.g., chemical
properties engine 630) may perform a selected machine learning rule
or algorithm using the training set, as described herein. Once
trained, the model may be used to determine (e.g., predict) the
identity and/or values of properties of the chemical composition,
relative to the property of interest.
[0129] FIG. 9A shows a block diagram of example data used for
training chemical properties engine 630. Data 902 may relate to one
or more chemical compositions (e.g., sample chemical compositions).
Data 902 may be known and/or determined. For example, data 902 may
be known by experimentally measuring the data, mathematically
calculating (e.g., via thermodynamic calculations) the data,
receiving the data from storage (e.g., from a database, such as a
chemical properties database 624), receiving the data from clinical
trials, etc. Data 902 may include values for one or more
parameters. For example, data 902 may include identities of
chemical compositions. Identities may include the names of the
chemical compositions, ingredients of the chemical compositions,
chemoinformatic values/properties of the ingredients of the
chemical compositions, values of properties of the chemical
compositions, consumer perceptions of the chemical compositions,
etc. Data may be input into the chemical properties engine 630, for
example, to train the model to predict one or more values of
chemical compositions.
[0130] As shown in FIG. 9A, data 902 may include one or more
ingredients 912 of one or more chemical compositions. Ingredients
may include a first ingredient 912a, a second ingredient 912b, etc.
Example ingredients are provided in FIGS. 1A, 1B and 2A, 2B. For
example, a chemical composition may include water, glycerin,
propylene glycol, and flavor ingredients. In such example, data 902
may include data for water, glycerin, propylene glycol, and flavor
ingredients. Each ingredient 912a, 912b, etc., may include an
identity of the ingredient and/or a chemoinformatic value for the
ingredient. For example, data 902 may include chemoinformatic
values 914a, 914b, etc. In the example chemical composition that
includes glycerin, propylene glycol, and flavor ingredients, each
of glycerin, propylene glycol, and flavor will have a respective
chemoinformatic value within data 902.
[0131] Data 902 may include a value 904 of a property of the
chemical composition. The value 904 of the property may be affected
by one or more of the ingredients. For example, the value 904 of
the property may be affected by one or more of the ingredients
interacting with one or more other ingredients of the chemical
composition. A property may be pH, fluoride stability, viscosity
stability, abrasion, specific gravity, a consumer perception of the
chemical composition, etc. Data 902 may include a value of the
property, such as a value of a pH property. As described herein,
the value of the property (e.g., pH property) may be affected by
one or more of the ingredients of the chemical composition.
[0132] One or more values of data 902 may be input into the
chemical properties engine 630, for example, to train the chemical
properties engine 630. Identities of chemical compositions and
associated values of properties (e.g., other properties) of the
chemical compositions may be input into the chemical properties
engine 630. As an example, ingredients of chemical composition
(e.g., sample chemical composition) and associated values of
parameters may be input into the chemical properties engine 630.
The chemical properties engine 630 may provide an association of
the ingredients of the chemical compositions and the values of
parameters of the chemical compositions.
[0133] FIG. 9B shows a block diagram of example data 920 used for
determining information relating to a chemical composition using
chemical properties engine 630. For example, FIG. 9B shows a block
diagram of example data 920 used for determining a value of a
property via chemical properties engine 630. Data 920 may relate to
one or more chemical compositions (e.g., considered chemical
compositions). Data 920 may be known and/or determined. For
example, data 920 may be known and/or determined by receiving the
data from storage (e.g., from a database, such as a chemical
properties database 624), experimentally measuring the data,
mathematically calculating (e.g., via thermodynamic calculations)
the data, etc. Data 920 may include values for one or more
parameters. Data 920 may include identities of chemical
compositions, such as the names of the chemical compositions,
ingredients of the chemical compositions, chemoinformatic
values/properties of the ingredients of the chemical compositions,
values of properties of the chemical compositions, etc. Data 920
may be input into the chemical properties engine 630, for example,
to determine from the chemical properties engine 630 (e.g., machine
learning model of chemical properties engine 630) one or more
values of properties of chemical compositions.
[0134] Data 920 may include one or more ingredients 912 of one or
more chemical compositions. Ingredients may include a first
ingredient 912a, a second ingredient 912b, etc. For example, a
chemical composition may include water, glycerin, propylene glycol,
and flavor ingredients. In such example, data 920 may include data
for water, glycerin, propylene glycol, and flavor ingredients. Each
ingredient 912a, 912b, etc., may include a chemoinformatic value
for the ingredient. For example, data 902 may include
chemoinformatic values 914a, 914b, etc. In the example chemical
composition that includes glycerin, propylene glycol, and flavor
ingredients, each of glycerin, propylene glycol, and flavor will
have a respective chemoinformatic value within data 920.
[0135] One or more values of data 902 may be input into the
chemical properties engine 630, for example, to determine (e.g.,
determine from the chemical properties engine 630) a value of a
property of the chemical composition. For example, ingredients of a
chemical composition (e.g., a sample chemical composition) may be
input into the chemical properties engine 630. The chemical
properties engine 630 may run (e.g., process) one or more machine
learning rules to determine a value of a property of the chemical
composition. The chemical properties engine 630 may provide the
value of the property of the chemical composition after determining
the value.
[0136] Chemical properties engine 630 (e.g., model of chemical
properties engine 630) may be configured to predict a value of a
property of the chemical composition, for example, based on
receiving identities of ingredients of a chemical composition, etc.
In an example, the values of the properties of the chemical
composition may be affected by ingredients of the chemical
composition (e.g., may be affected by an interaction of one or more
of the ingredients of the chemical composition). When information
relating to a chemical composition is supplied to the trained
model, the output may comprise a prediction regarding the value of
the property of the chemical composition, a fitting parameter
associated with the chemical composition, identities of the
chemical composition, etc. The property may relate to a pH of the
chemical composition, a viscosity stability of the chemical
composition, an abrasion of the chemical composition, a specific
gravity of the chemical composition, a consumer perception of the
chemical composition, etc. The predictions may take the form of a
value from a continuous range of values or from a discrete value,
for example.
[0137] FIG. 9C shows a block diagram of example data 930 used for
determining an identity of a chemical composition via chemical
properties engine 630. Data 930 may relate to one or more chemical
compositions (e.g., considered chemical compositions). Data 930 may
be known. For example, as described herein, data 930 may be known
by experimentally measuring the data, mathematically calculating
(e.g., via thermodynamic calculations) the data, receiving the data
via a clinical consumer trial, receiving the data from storage
(e.g., from a database, such as a chemical properties database
624), etc. As shown in FIG. 9C, data 930 may include a value 904 of
a property of a chemical composition, etc. Data 930 may be input
into the chemical properties engine 630, for example, to determine
information (e.g., associated information) of chemical
compositions. For example, data 930 (e.g., value 904) may be input
into the chemical properties engine 630, for example, to determine
from the model ingredients 912 of the chemical composition
determined (e.g., predicted) to relate to the value 904 of the
property input into the chemical properties engine 630.
[0138] As described herein, data 930 may include a value 904 of a
property. The value 904 of the property may be input into chemical
properties engine 630, for example, to predict (e.g., determine) a
name of a chemical composition, one or more ingredients 912 of a
chemical composition, chemoinformatic values 914a, 914b, . . . ,
914n of a chemical composition, etc. Ingredients may include a
first ingredient 912a, a second ingredient 912b, etc. For example,
a chemical composition may include water, glycerin, propylene
glycol, and flavor ingredients. A chemoinformatic value may be
associated with an (e.g., each) ingredient 912a, 912b, etc. For
example, ingredient 912a may include chemoinformatic value
914a.
[0139] As an example, a value 904 of a property of the chemical
composition may be input into chemical properties engine 630. Based
on value 904 of the property of the chemical composition, the
chemical properties engine 630 (e.g., model of chemical properties
engine 630) may be configured to predict identities (e.g., names,
ingredients, chemoinformatic values, etc.) of a chemical
composition forming a personal care product. When information
relating to a chemical composition is supplied to the chemical
properties engine 630, the output may comprise a determination
(e.g., prediction) regarding the identities (e.g., names,
chemoinformatic values, etc.) of a chemical composition forming the
personal care product, a fitting parameter associated with the
chemical composition, a value (e.g., another value) of a property
of the chemical composition, etc. The chemical properties engine
630 may provide the names, ingredients, chemoinformatic values,
etc. of the chemical composition to user via user device 502, for
example.
[0140] FIGS. 10A-10C show example graphical user interfaces (GUIs)
for training a chemical properties modeling device 602 (e.g., a
chemical properties engine 630 within chemical properties modeling
device 602). The GUI may be displayed on one or more devices. For
example, the GUI may be displayed on a training device, such as
training device 650, a user device, etc.
[0141] As shown in FIG. 10A, the GUI may request information from
the user, for example, the GUI may request information from the
user via prompt request 1010. Prompt request 1010 may ask the user
what data the user would like to use to train the chemical
properties engine 630 (e.g., model of chemical properties engine
630). The data used to train the model may be referred to as sample
data. The data may include an identity (e.g., name, ingredients,
chemoinformatic values of ingredients) of a chemical composition
forming a personal care product, a value of a property of personal
care product, etc. The GUI may provide an input mechanism 1012 for
the user to provide a response to prompt request 1010. For example,
the GUI may have a text box for receiving text from the user, a
radio button for selection, etc. As shown in FIG. 10A, check box
1012 may be provided. In examples in which a text box is provided,
the user may check one or more of the data in the input mechanism
1012 for training the chemical properties engine 630.
[0142] After the user selects the data desired to be input into the
chemical properties engine 630 (e.g., for training the chemical
properties engine 630), the user may input such data. The user may
input the data manually (e.g., via manually typing or speaking the
data). The user may input a single piece of data or the user may
input multiple pieces of data. For example, as shown in FIG. 10B,
GUI may provide an indication for the user to select the data to be
input into the chemical properties engine 630. The GUI may display
the data to be input into the chemical properties engine 630, For
example, the GUI may display the data to be input into the chemical
properties engine 630 based on the input provided on input
mechanism 1012 of FIG. 10A.
[0143] In an example, the user may desire to input ingredients
1016a of a chemical composition forming a personal care product and
a value 1016b of a property of the chemical composition. As shown
in FIG. 10B, the user may select a file to provide the ingredient
information (via Browse 1017a) and the value (via Browse 1017b) of
the property information. Although FIG. 10B shows a Browse button
for inputting data, one of skill in the art will understand that
other methods exist for selecting and/or inputting data into
chemical properties engine 630, such as via a database (such as a
database housed on a server, such as a cloud server), via one or
more hard drives, via external devices (such as user device 502),
etc.
[0144] The user may input data into the chemical properties engine
630 for one or more chemical compositions. For example, the user
may train the chemical properties engine 630 with data relating to
tens, hundreds, thousands, etc., of chemical compositions. The user
may train the chemical properties engine 630 with the same data for
one or more of the chemical compositions. For example, the user may
train the chemical properties engine 630 with ingredients and
values of properties of dozens of chemical compositions.
[0145] The user may train the chemical properties engine 630 with
different data (e.g., types of data) for one or more of the
chemical compositions. For example, the user may train the chemical
properties engine 630 with ingredients and values of properties for
some of the chemical compositions, with chemoinformatic values and
values of properties for some of the chemical compositions, with
names of chemical compositions and values of properties for some of
the chemical compositions, etc. The user may train the chemical
properties engine 630 with values of properties comprising a pH, a
fluoride stability, a viscosity stability, an abrasion, a specific
gravity, consumer perceptions, etc. The training device 650 (e.g.,
GUI of training device) may request if the user desires to input
additional data. For example, as shown in FIG. 10C, the GUI may
provide an additional data prompt 1018 asking the user if the user
desires to input any additional data into the chemical properties
engine 630 (e.g., the model of chemical properties engine 630). The
user may desire to input additional data into chemical properties
engine 630 if the user desires to further train the chemical
properties engine 630. If the user desires to input additional data
into the chemical properties engine 630, the user may select the
Yes prompt in area 1020, otherwise the user may select the No
prompt in area 1020. If the user selects the Yes prompt in area
1020, the GUI shown in FIG. B (and described herein) may be
provided to the user. If the user selects the No prompt in area
1020, the user may exit the GUI.
[0146] FIGS. 11A-11D show example graphical user interfaces (GUIs)
for determining (e.g., predicting) data from a chemical properties
modeling device 602 (e.g., a chemical properties engine 630 within
chemical properties modeling device 602). The GUI may be displayed
on one or more devices. For example, the GUI may be displayed on a
user device, such as user device 502.
[0147] As shown in FIG. 11A, the GUI may request information from
the user, for example, via prompt request 1110. Prompt request 1110
may ask the user what data the user would like the model to
determine (e.g., predict). The GUI may provide an input mechanism
1112 for the user to provide a response to prompt request 1110. For
example, the GUI may have a text box for receiving text from the
user, a radio button for selection, etc. As shown in FIG. 11A,
check box 1112 may be provided. The user may check one or more of
the data in the input mechanism 1112 so that the chemical
properties engine 630 may determine one or more pieces of data
relating to the chemical composition. Input mechanism may permit
additional information, including sub-categories of information, to
be selected for determination. For example, input mechanism 1112
may allow a user to further define the value of the property to be
determined to be one a pH, a fluoride stability, a viscosity
stability, an abrasion, a specific gravity, a consumer perception,
etc.
[0148] After the user selects what data the user desires the
chemical properties engine 630 to determine, the user may input
data that is associated with the desired data, as shown in FIG.
11B. For example, prompt 1116 indicates that the user desires to
determine a value of a property of the chemical composition (based
on the user's input at input 1112, in FIG. 11A). As shown in FIG.
11B, the GUI may provide input 1118, allowing the user to select
what data the user desires to input into the chemical properties
engine 630, for example, to determine the value of the property of
the chemical composition. Examples of data to be input into the
chemical properties engine 630 includes identity (e.g., name,
ingredients, chemoinformatic values of ingredients) data of a
chemical composition forming a personal care product, a value of a
property of personal care product, etc.
[0149] After the user selects the data that the user desires to
determine (1112), and the data that the user would like to use to
determine the value of the personal care product (1118), the user
may provide the associated data. FIG. 11C shows an example GUI in
which user may at input data, at 1122. Prompt 1120 indicates that
the user has chosen to input ingredients of the chemical
composition (to determine a value of a property of the chemical
composition), however such indication is for illustration purposes
only and other types of data may be provided by user to determine
information relating to chemical composition.
[0150] The user may input the data manually (e.g., via manually
typing or speaking the data). For example, the user may input the
ingredients of the chemical composition, as shown on FIG. 11C,
manually. The user may input a single piece of data or the user may
input multiple pieces of data. For example, as shown in FIG. 11C,
GUI may provide an indication for the user to select the data to be
input into the chemical properties engine 630. The user may select
a file to provide the ingredient information (via Browse button
1122). Although FIG. 11C shows a Browse button 1122 for inputting
data, one of skill in the art will understand that other methods
exist for selecting and/or inputting data into chemical properties
engine 630, such as via a database (such as a database housed on
the cloud), via external hard drives, via external devices (such as
user device 502), etc.
[0151] The GUI may provide the determined (e.g., predicted) data.
For example, the GUI may provide the value of a parameter, as shown
in FIG. 11D. The value of the parameter may relate to a chemical
composition forming a personal care product. Prompt 1130 may
display the associated data provided by the user. For example,
prompt 1130 may display that the determined data is based on the
ingredient information (e.g., the ingredient information provided
by the user). Prompt 1130 may indicate what data has been
determined. For example, prompt 1130 indicates that the value of
the property of the personal care product is being determined.
Output 1132 provides the determined value. As shown on FIG. 11D,
the determined value may be 1.7. In examples the GUI may provide
further information of the information such as that the property is
a pH, a fluoride stability, a viscosity stability, an abrasion, a
specific gravity, and/or a consumer perception.
[0152] FIG. 12 is an example process 1200 for determining (e.g.,
predicting) a value of a chemical composition. The value may be an
identity of a chemical composition, such as a name of the chemical
composition, ingredients of the chemical composition,
chemoinformatic values of ingredients of the chemical composition,
a value of a property of the chemical composition, etc.
[0153] At 1202, an identity of a chemical composition (e.g., a
sample chemical composition) may be received. The identity may be
received from a database or another storage device, a described
herein. As provided above, the identity of the chemical composition
may be a name of the chemical composition, ingredients of the
chemical composition, chemoinformatic values of ingredients of the
chemical composition, etc.
[0154] At 1204, a value of a parameter of the chemical composition
(e.g., a sample chemical composition) may be received. The value of
the property may be affected by one or more ingredients of the
chemical composition. As an example, the property may be a pH of
the chemical composition, and the value of the property may be the
value of the pH of the chemical composition.
[0155] The identity of the chemical composition (e.g., the sample
chemical composition) and/or the value of the parameter may be used
to train a machine learning model, as described herein. For
example, at 1206 the value of the property of the chemical
composition (e.g., the sample chemical composition) may be input
into the machine learning model to train the machine learning
model. An identity of the chemical composition may be input into
the machine learning model to train the machine learning model. The
identity of the chemical composition may be one or more of a name
of the chemical composition, ingredients of the chemical
composition, chemoinformatic values of ingredients of the chemical
composition, etc. The machine learning model may make associations
of the value of the property of the chemical composition and the
identity of the chemical composition.
[0156] After the machine learning model is trained, the machine
learning model may determine one or more values of a chemical
composition. The machine learning model may determine one or more
values of a chemical composition in response to receiving an
associated piece of data. For example, the machine learning model
may determine a value of a property of a chemical composition based
on an identity of the chemical composition, such as ingredients of
the chemical composition or a name of the chemical composition. The
property of the chemical composition may be a pH value of the
chemical composition, a fluoride stability value of the chemical
composition, a viscosity value of the chemical composition, an
abrasion value of the chemical composition, a specific gravity
value of the chemical composition, a consumer perception value of a
chemical composition, etc.
[0157] For example, at 1208 the machine learning model may receive
one or more values of a chemical composition (e.g., a considered
chemical composition). As described herein, a considered chemical
composition may be a chemical composition in which one or more
values are unknown and desired to be known. For example, an
identity of a considered chemical composition may be known,
ingredients of the considered chemical composition may be known,
and/or chemoinformatic values of the considered chemical
composition may be known. The value of a property of the considered
chemical composition may be unknown.
[0158] The known values (e.g., the identity of the considered
chemical composition, ingredients of the considered chemical
composition, and/or chemoinformatic values of the considered
chemical composition) may be input into the machine learning model.
Based on the values input into the machine learning model, the
machine learning model may determine a value of a property of the
considered chemical composition. The value of the considered
chemical composition may be displayed or otherwise provided to a
user.
[0159] In examples, the user may determine whether the value of the
property corresponds to a desired value of the property. For
example, it may be desired (e.g., required) that a personal care
product have a defined value for a property of the personal care
product. The value may relate to a pH value or one or more other
properties described herein. If the machine learning model
determines that a chemical composition has a value of the property
that aligns with a desired value of the property, the user may
perform an action, such as producing a personal care product having
the ingredients associated with the determined value. In other
examples, the user may perform actions to confirm that the results
provided by the machine learning model are accurate, such as by
performing a measurement of the value of the property, performing
mathematically calculations of the value. The user may confirm that
the results provided by the machine learning model are accurate
before producing a personal care product having the ingredients
associated with the determined value.
[0160] Systems described herein may be implemented using any
available computer system and adaptations contemplated for known
and later developed computing platforms and hardware. Further, the
methods described herein may be carried out by software
applications configured to execute on computer systems ranging from
single-user workstations, client server networks, large distributed
systems employing peer-to-peer techniques, or clustered grid
systems. In an example, a high-speed computing cluster may be used.
The computer systems used to practice the methods described herein
may be geographically dispersed across local or national boundaries
using a data communications network such as the Internet. Moreover,
predictions generated at one location may be transported to other
locations using well known data storage and transmission
techniques, and predictions may be verified experimentally at the
other locations.
[0161] While the invention has been described with respect to
specific examples including presently preferred modes of carrying
out the invention, those skilled in the art will appreciate that
there are numerous variations and permutations of the above
described systems and techniques. It is to be understood that other
embodiments may be utilized and structural and functional
modifications may be made without departing from the scope of the
present invention. Thus, the spirit and scope of the invention
should be construed broadly as set forth in the appended
claims.
* * * * *