ChatGPT Chemistry Assistant for Text Mining and the Prediction of MOF Synthesis

Abstract

We use prompt engineering to guide ChatGPT in the automation of text mining of metal–organic framework (MOF) synthesis conditions from diverse formats and styles of the scientific literature. This effectively mitigates ChatGPT’s tendency to hallucinate information, an issue that previously made the use of large language models (LLMs) in scientific fields challenging. Our approach involves the development of a workflow implementing three different processes for text mining, programmed by ChatGPT itself. All of them enable parsing, searching, filtering, classification, summarization, and data unification with different trade-offs among labor, speed, and accuracy. We deploy this system to extract 26 257 distinct synthesis parameters pertaining to approximately 800 MOFs sourced from peer-reviewed research articles. This process incorporates our ChemPrompt Engineering strategy to instruct ChatGPT in text mining, resulting in impressive precision, recall, and F1 scores of 90–99%. Furthermore, with the data set built by text mining, we constructed a machine-learning model with over 87% accuracy in predicting MOF experimental crystallization outcomes and preliminarily identifying important factors in MOF crystallization. We also developed a reliable data-grounded MOF chatbot to answer questions about chemical reactions and synthesis procedures. Given that the process of using ChatGPT reliably mines and tabulates diverse MOF synthesis information in a unified format while using only narrative language requiring no coding expertise, we anticipate that our ChatGPT Chemistry Assistant will be very useful across various other chemistry subdisciplines.

Realizing the cooking recipe of materials synthesis through large language models

“…By quantitatively measuring the similarity between LLM-output and the original synthesis information, we show that LLMs accurately structure the full process of synthesis in an ordered format, thereby enabling the subsequent analysis and learning beyond the prediction of individual synthesis parameters. Together with recent progress of utilizing LLMs in materials research, 31,32 our work further highlights the capability of LLMs in comprehending knowledge directly from materials literature and their massive potential in designing and guiding materials experimentation.…”

“…27–30 Their remarkable success in a wide variety of tasks has promoted significant interest in utilizing LLMs in the materials research domain. 31,32 For instance, the capability of LLMs to process text corpora allowed the identification and extraction of key parameters for materials synthesis without necessary human labeling, 31 thereby opening up the possibility of LLM-guided experimentation. 31,32…”

“…31,32 For instance, the capability of LLMs to process text corpora allowed the identification and extraction of key parameters for materials synthesis without necessary human labeling, 31 thereby opening up the possibility of LLM-guided experimentation. 31,32…”

Shaping the Water-Harvesting Behavior of Metal–Organic Frameworks Aided by Fine-Tuned GPT Models

“…Each category encompasses two to five specific mutation actions. Our journey began with over 200 linkers, extracted from the literature using a ChatGPT-based assistant we previously reported . From this, we concentrated on 142 carboxylate linkers, which are more commonly seen in MOF synthesis.…”

Can ChatGPT and Other AI Bots Serve as Peer Reviewers?

“…The recent emergence of artificial intelligence (AI) bots, such as ChatGPT, Bing AI Chat and Copilot (Microsoft), and Bard (Google), has led the scientific community to explore the capabilities of Large Language Models (LLMs) in scientific research . Recently, several editorials, articles, and blog posts have appeared discussing the pros and cons of using such AI bots in scientific publications. − On the positive side, these tools are useful for grammar correction and revision of text, to spark new ideas and break mental logjams, and for content analysis of published articles. These features are helpful for researchers to compose scientific articles, especially when English is not their first language of study.…”