PExAI a.k.a. a new research project on AI's role in chemistry has just started!
- Jargonium editors
- Apr 29
- 1 min read
Updated: Apr 30
In February 2026, our very own co-editor Vanessa Seifert commenced a new research project, based at the University of Athens, and called PExAI: Predicting and Explaining with AI. The project is the first of its kind, in the sense of briging together the philosophy of chemistry with the philosophy of AI!
The project will last two years and is funded by the Hellenic Foundation for Research and Innovation.
The project is set to investigate how AI is used in chemistry for the prediction of new chemical substances and the explaination of chemical reactions. The goal is to understand how AI contributes to chemical progress. The main case study is IMPRESSION, an AI model that is currently developed at the University of Bristol by the research team led by Prof. Craig Butts.
To celebrate the commencement of PExAI project, Vanessa is organising on the 18th of May a hybrid event with guest speakers Stathis Psillos, Craig Butts and Emanuele Ratti.
You can find all the details about the project and the upcoming event here: https://www.vanessa-seifert.com/event-details/kick-off-event
The event is open to all!! Webex link is this: https://uoa.webex.com/uoa/j.php?MTID=m2eb2a017c5bdbff6c6bd91e994e2b995
Excited to see PExAI tackling how AI can actually accelerate chemical discovery—curious if you'll be addressing reproducibility issues with ML models in synthesis planning. I've been digging into related tools lately and found something worth checking out. https://flux3.site
Excited to see PExAI tackling the reproducibility gap in computational chemistry—curious how you're planning to handle the "black box" problem with neural network potentials. I've been digging into similar workflows over at https://spheroz.com
Excited to see PExAI tackling the reproducibility problem in computational chemistry—that’s been the bottleneck in my own workflow for years. I’ve been digging into similar open-source tools lately, check out https://image-gpt.net
Excited to see PExAI tackling the reproducibility gap in computational chemistry—that's been the elephant in the room for too long. I've been digging into similar workflow automation tools myself, check out https://framepack-ai.com
Excited to see how PExAI tackles the reproducibility problem in computational chemistry—that’s been the bottleneck in my own workflow for years. Curious if you’ll be benchmarking against existing ML potentials, and I’ve been digging into related tools over at https://qwenimaging.com