In a recent presentation, researchers from Imperial College London detailed how combining ontological knowledge with argumentative agentic AI can significantly accelerate chemical development. Professor Alexei Lapkin, Professor Francesca Toni, and Dr Antonio Rago discussed their approach during a talk hosted by the AIchemyhubUK channel.
The method leverages structured ontological knowledge—a formal representation of concepts and their relationships within chemistry—to provide a foundation for AI reasoning. On top of this, argumentative agentic AI enables multiple AI agents to debate and justify decisions, simulating scientific peer review to refine hypotheses and experimental plans.
By integrating these two AI paradigms, the team aims to reduce the time required for chemical discovery and optimization. The agents can argue over reaction pathways, predict outcomes, and suggest novel experiments, all while maintaining a transparent chain of reasoning.
The presenters emphasized that this approach is not limited to chemistry but could be adapted to other scientific domains requiring complex knowledge integration and decision-making.
For more details, the full video is available on YouTube.