Rapid Material Property Prediction by Artificial Intelligence Lab's Assistant
In the ever-evolving world of science and technology, a new tool is making waves in the field of materials science. Kamal Choudhary, a professor at Johns Hopkins University's Whiting School of Engineering, has developed a groundbreaking AI tool called the ChatGPT Materials Explorer (CME).
The ultimate goal of CME is to serve as a one-stop research assistant for materials scientists, aiding them in computer simulations, data analysis, and other methods that advance the field. CME is designed to assist researchers by accessing real scientific data and physics-based models.
CME is connected to several key databases, including the National Institute of Science and Technology-Joint Automated Repository for Various Integrated Simulations (NIST-JARVIS), National Institutes of Health-Chemistry Agent Connecting Tool Usage to Science (NIH-CACTUS), and the Materials Project. These databases ensure that CME is always updated with the most recent materials science findings.
While data sources like Wikipedia or The New York Times don't often include current facts and research about materials science, CME is designed to provide accurate answers. However, it's important to note that, as with any AI, CME is not infallible. Choudhary found that, in some cases, CME may present false information, a phenomenon known as hallucinations.
To combat this issue, Choudhary developed his specialized language model with the ChatGPT builder feature, connecting the AI to the databases and instructing it on what kinds of answers it can give. When it can't find the exact answer based on the data it's pulling from, it will say something that sounds plausible, but may not be accurate.
In testing, CME proved to be a valuable resource. When posed with eight chemistry-related tasks, CME got all eight answers correct, while ChatGPT 4 and ChemCrow only gave five accurate responses.
Choudhary's open-source platform, Atom GPT, offers a contrast to the closed-source model of CME. While users can't edit the code that Choudhary established in CME, Atom GPT allows select users to change the code and improve its ability to answer materials science questions. AtomGPT.org, the home of Atom GPT, was developed by a group of independent developers and contributors; no single individual is publicly credited as the sole creator.
Hallucinations occur because ChatGPT isn't trained to understand facts. Choudhary is working to develop the platform further by adding advanced materials modeling tools, automated literature reviews, and more. He hopes that these improvements will further reduce the occurrence of hallucinations and make CME an even more valuable tool for materials scientists.
The project is tagged under both materials science and artificial intelligence, reflecting its dual nature. As research in these fields continues to advance, it's clear that tools like CME will play an increasingly important role in pushing the boundaries of what's possible.
This article is posted in the Science+Technology category.
Read also:
- AI-Powered X-Nave Platform and Fresh Gaming Content to be Demonstrated by EGT Digital at SBC Summit Lisbon Event
- British technology company Nvidia invests a vast sum of £11 billion in AI technology within the U.K., announcing this during a visit by U.S. President Trump.
- Rapid advancement of AI technology poses potential threat to job stability, according to AI CEO's remarks.
- Spheron and Nubila Team Up to Use Web3 Technology for AI that Combats Climate Change
