CAMEL AI Live Talk: Biomni by Kexin Huang

At a recent CAMEL AI event, Kexin Huang, who presented two groundbreaking systems at the intersection of AI and biomedical research. 🧬 First up was Biomni – a general biomedical AI agent designed to automate a wide range of scientific tasks, from retrieval and reasoning to code generation. 🔬 Biomni has demonstrated near human-level performance in complex challenges like cloning protocol generation and large-scale data analysis. 📊 These results were validated through collaborations with Professor Leon and Michael Snyder's group, and Biomni is now available through a web platform for scientists worldwide. 🧪 Then came Proper – a multi-agent system inspired by falsification theory. It enables rigorous hypothesis testing with built-in statistical checks, achieving results on par with human experts – but much faster.