ÌìÃÀ´«Ã½

An Alchemy Frontier Fund project involving ÌìÃÀ´«Ã½ will develop explainable, argumentative AI agents that combine reinforcement learning and expert knowledge, supporting human decision making in complex pharmaceutical manufacturing and chemical process development contexts.

from the Department of Computing at ÌìÃÀ´«Ã½ has secured £800,000 from the alongside Professor Alexei Lapkin at the University of Cambridge and Dr Antonio Rago at King’s College London. The project is supported by industry partners and the startup and is one of only three selected from 50 applications. 

The project, titled Leveraging Ontological Knowledge with Argumentative Agentic AI to Accelerate Chemical Development, will develop specialised artificial intelligence (AI) agents designed to work alongside human experts. These agents will support established computational chemistry approaches used in drug development. They will use reinforcement learning combined with computational argumentation and expert ontological knowledge of chemical processes to reach decisions in an explainable way, operating alongside humans and other agents in multi‑agent systems able to engage in argumentative exchanges. The agents will be tested in an automated simulation environment developed at the University of Cambridge. 

This is an amazing opportunity to bring together research in computational argumentation, reinforcement learning and multi-agent systems to bear impact in chemical development settings." Professor Francesca Toni

AI is increasingly being applied across discovery science, engineering and manufacturing, with AI for chemistry emerging as a key discipline. However, the complexity of chemical development and pharmaceutical manufacturing, combined with limited access to high quality data, continues to make the translation of research into commercial practice challenging. This project addresses these issues through the development of explainable, human in the loop AI systems that integrate computational argumentation, reinforcement learning and expert knowledge to support decision making in real world settings. 

Commenting on the funded project, Professor Francesca Toni said: "This is an amazing opportunity to bring together research in computational argumentation, reinforcement learning and multi-agent systems to bear impact in chemical development settings, in collaboration with two additional top institutions, industry and the Alchemy hub. I am super excited at the prospect."  

Over the two-year lifetime of the project, the team aims to establish a robust pipeline for applying this approach within pharmaceutical manufacturing. They also plan to explore its use in other areas of chemical science involved in the manufacture and processing of materials.