Seeking fellow computational chemists/physicists!
I am currently working full-time as an engineer at a major silicon company, so this would be an asynchronously updated weekend project.
The idea is to train an LLM with time series data and let it predict the next time iteration. We're going to basically treat each time iteration as a "word", not actually treating them as numbers . LLMs are notoriously bad at math so we're going to be avoiding that altogether.
This is primarily a self education project and I'm not looking to break new ground or anything, I have a few publications on very similar stuff done before.
Any takers?