Kevin Yager, a researcher at the Center for Functional Nanomaterials (CFN), has created a chatbot with a knowledge of nanomaterial science using new advances in artificial intelligence (AI) and machine learning (ML). This specialized AI chatbot can help scientists brainstorm and ideate their research. The bot uses language models and text retrieval methods to accurately understand and answer complex scientific questions. It is designed to efficiently summarize larger pieces of work and transcend language barriers.
Rise of the Robots
The chatbot is part of a larger effort at the CFN to use AI/ML to accelerate the discovery of new nanomaterials and automate experiments, which is already proving useful at the National Synchrotron Light Source II (NSLS-II). By leveraging scientific text, the bot can quickly understand and convey information about various experiments and help with scientific communication, saving time and ensuring accurate information is being conveyed
Watching Your Language
To create the specialized chatbot, Kevin Yager’s team used domain-specific text from scientific publications to train the AI model, which help the bot to be more accurate and trustworthy in its responses. The team developed methods to prevent the bot for creating false information or citations. They used a process called “embedding” to quickly and accurately categorize and link scientific information behind the scenes. This method also enables the bot to understand the context of the user’s question, ensuring the text is relevant and the answers are factual and sourced.
Bots Empowering Humans
The AI/ML systems that CFN is developing are designed to free up human researchers’ time by automating repetitive tasks and making scientific research more efficient. The domain-specific chatbot can help scientists organize, summarize, and find relevant information in scientific documents, effectively accelerating the research process.
In Conclusion
The use of AI and ML in scientific research is raising important discussions about safety and ethics. Developing these AI assistants holds great promise for accelerating the scientific discovery process, and the potential applications for these bots in research are vast. The CFN is excited to see where this technology will go and is looking forward to what will be possible in the future.