# US 2025/0095796 A1 · Methods of designing polymers and polymers designed therefrom

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0095796 A1
- Document ID: US20250095796A1
- Status: Published application
- Kind: A1
- Application number: 18/794,295
- Filed: 2024-08-05
- Published: 2025-03-20
- Earliest claimed: 2023-09-15
- Assignees: Toyota Research Institute, Inc.; Toyota Jidosha Kabushiki Kaisha
- Applicants: Toyota Research Institute, Inc.
- Inventors: Arash Khajeh; Ha-Kyung Kwon; Daniel Schweigert; Zhenze Yang; Weike Ye; Xiangyun Lei
- Web page: https://patentgrep.com/patent/US20250095796A1

## CPC

- G16C 20/50: Molecular design, e.g. of drugs
- G16C 20/30: Prediction of properties of chemical compounds, compositions or mixtures
- G16C 20/70: Machine learning, data mining or chemometrics
- G16C 60/00: Computational materials science, i.e. ICT specially adapted for investigating the physical or chemical properties of materials or phenomena associated with their design, synthesis, processing, characterisation or utilisation
- G16C: COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE

## Abstract

A method for designing polymers includes translating polymer representations of a training dataset and a test dataset into a format comprehensible by a generative pretraining transformer (GPT)-based model, training the GPT-based model with the translated polymer representations, generating new polymer representations, in a predefined format, using the trained GPT-based model, predicting at least one property of the generated new polymer representations using a machine learning (ML) property predictive model and selecting a first subset of the generated new polymer representations as a function of the at least one predicted property, and calculating the at least one property of the first subset of the generated new polymer representations using a molecular dynamics (MD) module.

## Claims

### Claim 1

A method for designing polymers using a machine learning system, comprising:
translating polymer representations of a training dataset and a test dataset into a format comprehensible by a generative pretraining transformer (GPT)-based model;
training the GPT-based model with the translated polymer representations;
generating new polymer representations, in a predefined format, using the trained GPT-based model;
predicting at least one property of the generated new polymer representations using a machine learning (ML) property predictive model and selecting a first subset of the generated new polymer representations as a function of the at least one predicted property; and
calculating the at least one property of the first subset of the generated new polymer representations using a molecular dynamics (MD) module.

### Claim 2 (depends on 1)

The method according to claim 1 further comprising tokenizing the translated polymer representations before training the GPT-based model.

### Claim 3 (depends on 2)

The method according to claim 2, wherein the training dataset and the test dataset comprise representations of known monomers of polymer electrolytes.

### Claim 4 (depends on 3)

The method according to claim 3, wherein the training dataset and the test dataset further comprise corresponding property values for the known monomers.

### Claim 5 (depends on 4)

The method according to claim 4, wherein the corresponding property values are selected from the group consisting of ionic conductivity values, transference number values, and density values.

### Claim 6 (depends on 1)

The method according to claim 1, wherein the translated polymer representations are tokenized p-SMILES strings.

### Claim 7 (depends on 6)

The method according to claim 6, wherein the new polymer representations are p-SMILES strings generated by the trained GPT-based model.

### Claim 8 (depends on 7)

The method according to claim 7, wherein the generated p-SMILES strings comprise one or more selected from the group consisting of *ONCCOC*, *OCCOC*, *OCCOCCN*, *OCCOCCOC*, *COCC*, *OCCCOCCN*, *SCCOCCN*, *OCCCOC*, *OCCCON*, *SCCOC*, *OOCCOCCN*, *OCCCOCC*, *CCOCOCCN*, and *OCCOCCSC *.

### Claim 9 (depends on 1)

The method according to claim 1 further comprising selecting and providing a second subset of the generated new polymer representations to the training dataset or the test dataset such that training of the GPT-based model includes the second subset of the generated new polymer representations.

### Claim 10 (depends on 9)

The method according to claim 9, wherein the second subset of the generated new polymer representations are selected as a function of the at least one property of the first subset of the generated new polymer representations calculated using a molecular dynamics (MD) module.

### Claim 11

A system for designing polymers, the system comprising:
a processor and a memory communicably coupled to the processor and storing machine-readable instructions that, when executed by the processor, cause the processor to:
train a generative pretraining transformer (GPT)-based model with polymer representations;
generate new polymer representations, in a predefined format, using the trained GPT-based model;
predict at least one property of the generated new polymer representations using a machine learning (ML) property predictive model and selecting a first subset of the generated new polymer representations as a function of the at least one predicted property; and
calculate the at least one property of the first subset of the generated new polymer representations using a molecular dynamics (MD) module.

### Claim 12 (depends on 11)

The system according to claim 11, wherein the memory communicably coupled to the processor and storing machine-readable instructions that, when executed by the processor, further cause the processor to train the GPT-based model with tokenized polymer representations.

### Claim 13 (depends on 12)

The system according to claim 12, wherein the training dataset and the test dataset comprise representations of known monomers of polymer electrolytes.

### Claim 14 (depends on 13)

The system according to claim 13, wherein the training dataset and the test dataset further comprise corresponding property values for the known monomers.

### Claim 15 (depends on 14)

The system according to claim 14, wherein the corresponding property values are selected from the group consisting of ionic conductivity values, transference number values, and density values.

### Claim 16 (depends on 11)

The system according to claim 11, wherein the polymer representations in the training dataset and the test dataset are in a p-SMILES format that is comprehensible by the GPT-based model.

### Claim 17 (depends on 16)

The system according to claim 16 further comprising tokenizing the p-SMILES format of the polymer representations before training the GPT-based model.

### Claim 18 (depends on 17)

The system according to claim 17, wherein the new polymer representations are one or more p-SMILES selected from the group consisting of *ONCCOC*, *OCCOC*, *OCCOCCN*, *OCCOCCOC*, *COCC*, *OCCCOCCN*, *SCCOCCN*, *OCCCOC*, *OCCCON*, *SCCOC*, *OOCCOCCN*, *OCCCOCC*, *CCOCOCCN*, and *OCCOCCSC *.

### Claim 19

A method for designing polymers using a machine learning system, comprising:
selecting a training dataset and a test dataset containing tokenized p-SMILES strings of known monomers of polymer electrolytes;
training the GPT-based model with the tokenized p-SMILES strings;
generating new polymer representations in a p-SMILES format using the trained GPT-based model;
predicting at least one property of the generated new polymer representations using a machine learning (ML) predictive property model and selecting a first subset of the generated new polymer representations as a function of the at least one predicted property; and
calculating the at least one property of the first subset of the generated new polymer representations using a molecular dynamics (MD) module.

### Claim 20 (depends on 19)

The method according to claim 19 further comprising selecting and providing a second subset of the generated new polymer representations to the training dataset or the test dataset such that training of the GPT-based model includes the second subset of the generated new polymer representations, wherein the second subset of the generated new polymer representations are selected as a function of the at least one property of the first subset of the generated new polymer representations calculated using a molecular dynamics (MD) module.
