Benchmarking large language models for materials synthesis: the case of atomic layer deposition
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913882261422080 |
|---|---|
| author | Yanguas-Gil, Angel Dearing, Matthew T. Elam, Jeffrey W. Jones, Jessica C. Kim, Sungjoon Mohammad, Adnan Nguyen, Chi Thang Sengupta, Bratin |
| author_facet | Yanguas-Gil, Angel Dearing, Matthew T. Elam, Jeffrey W. Jones, Jessica C. Kim, Sungjoon Mohammad, Adnan Nguyen, Chi Thang Sengupta, Bratin |
| contents | In this work we introduce an open-ended question benchmark, ALDbench, to evaluate the performance of large language models (LLMs) in materials synthesis, and in particular in the field of atomic layer deposition, a thin film growth technique used in energy applications and microelectronics. Our benchmark comprises questions with a level of difficulty ranging from graduate level to domain expert current with the state of the art in the field. Human experts reviewed the questions along the criteria of difficulty and specificity, and the model responses along four different criteria: overall quality, specificity, relevance, and accuracy. We ran this benchmark on an instance of OpenAI's GPT-4o. The responses from the model received a composite quality score of 3.7 on a 1 to 5 scale, consistent with a passing grade. However, 36% of the questions received at least one below average score. An in-depth analysis of the responses identified at least five instances of suspected hallucination. Finally, we observed statistically significant correlations between the difficulty of the question and the quality of the response, the difficulty of the question and the relevance of the response, and the specificity of the question and the accuracy of the response as graded by the human experts. This emphasizes the need to evaluate LLMs across multiple criteria beyond difficulty or accuracy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_10477 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Benchmarking large language models for materials synthesis: the case of atomic layer deposition Yanguas-Gil, Angel Dearing, Matthew T. Elam, Jeffrey W. Jones, Jessica C. Kim, Sungjoon Mohammad, Adnan Nguyen, Chi Thang Sengupta, Bratin Machine Learning Materials Science Artificial Intelligence In this work we introduce an open-ended question benchmark, ALDbench, to evaluate the performance of large language models (LLMs) in materials synthesis, and in particular in the field of atomic layer deposition, a thin film growth technique used in energy applications and microelectronics. Our benchmark comprises questions with a level of difficulty ranging from graduate level to domain expert current with the state of the art in the field. Human experts reviewed the questions along the criteria of difficulty and specificity, and the model responses along four different criteria: overall quality, specificity, relevance, and accuracy. We ran this benchmark on an instance of OpenAI's GPT-4o. The responses from the model received a composite quality score of 3.7 on a 1 to 5 scale, consistent with a passing grade. However, 36% of the questions received at least one below average score. An in-depth analysis of the responses identified at least five instances of suspected hallucination. Finally, we observed statistically significant correlations between the difficulty of the question and the quality of the response, the difficulty of the question and the relevance of the response, and the specificity of the question and the accuracy of the response as graded by the human experts. This emphasizes the need to evaluate LLMs across multiple criteria beyond difficulty or accuracy. |
| title | Benchmarking large language models for materials synthesis: the case of atomic layer deposition |
| topic | Machine Learning Materials Science Artificial Intelligence |
| url | https://arxiv.org/abs/2412.10477 |