Kamal Choudhary, a professor at Johns Hopkins University’s Whiting School of Engineering, has developed an innovative artificial intelligence tool named the ChatGPT Materials Explorer (CME). This specialized AI system, tailored specifically to the needs of materials scientists, can predict material properties in seconds. CME takes a few cues from the popular powers of ChatGPT for programmers and authors. Its aim is to speed up the discovery of new exotic materials, such as batteries and alloys.
Choudhary holds an appointment in the Department of Electrical and Computer Engineering. He began developing CME as a side project. Ultimately, the aggregators’ main purpose is to help materials scientists design new materials with optimal compositions and crystal structures. By leveraging CME, researchers can pose complex questions like, “Can you design a superconductor with a particular composition and show me the crystal structure?” The tool’s development was recently chronicled in the journal Integrating Materials and Manufacturing Innovation.
Designed for Efficiency
CME seeks to address potential biases typically found in traditional AI models. Kamal Choudhary focused on the issues of “hallucinations,” or when AI makes things up (in other words, when it generates accurate-sounding but incorrect or nonsensical answers).
“Hallucinations happen because ChatGPT isn’t trained to understand facts,” – Kamal Choudhary
He went into more detail on why using appropriate databases are essential when training AI systems.
“These databases are how ChatGPT gets its information, so plugging in databases that are relevant to the field is crucial to getting the correct output from the chatbot,” – Kamal Choudhary
CME has been filling this gap with a commitment to making sure it’s based on accurate, current information. The system is continually updated with the most recent research results. This best positions it to continue providing the most precise predictions and analyses for a variety of materials science questions.
Accelerating Material Discovery
If realized, the implementation of CME has the potential to massively accelerate the rate of material discovery. Academics who typically go through weeks of experiments to find the properties of materials can now, with CME, get the answers back in near real-time. This efficiency is particularly important in sectors such as energy storage and advanced manufacturing, where new ideas are leading rapid transformation.
This high-level accuracy enables researchers to spend more time on experimentation and development rather than processing data through a filtering process.
“Before, I would ask regular ChatGPT for the molecular structure notation of ibuprofen, and it would give an incorrect or generic response. With CME, I’ll get the right answer to this and many other materials science questions.” – Kamal Choudhary
Choudhary imagines CME as an all-encompassing research assistant designed specifically for materials scientists. Ideally, we’re looking to create a truly transformative tool. For one thing, it will improve research by bolstering everything from computer simulations to data analysis, scientific illustration, and writing.
The Future of Materials Science Research
Kicked up CME’s automation with always on updates from the latest journals which mean users always have access to the most up to date information.
“ChatGPT Materials Explorer is like having a specialized research assistant who is trained specifically to dig through huge databases, predict how a material or materials will behave without physical testing, sort through scientific papers to find studies relevant to your projects, and even analyze work and assist with scientific writing,” – Kamal Choudhary
CME enables materials scientists with cutting-edge predictive capabilities. If deployed wisely, this advance could fundamentally change how researchers approach the problem of material development.
“Materials Explorer is correct because these databases are automatically updated with new papers; it runs itself and pulls from the newest journals,” – Kamal Choudhary
By enhancing the capabilities of materials scientists through its advanced predictive features, CME stands to revolutionize how researchers approach material development.