I was thinking about Marianne's question regarding compression and MaxEnt models fit to data and Greg's recent paper optimizing-reservoir-computing-for-reconstructing-ergodic-properties where they constrained based on matching the invariant distribution. Seems like a connection to explore here
MaxEnt <-> Invariant Distribution
Comments (0)
No comments yet.
Most recent thought
May 25, 2026
Compression
On compression, I’ve recently been wondering if it makes sense to argue for fitting a MaxEnt model to data and then classifying based on the model rather based on individual features.
Most recent reference
May 26, 2026
Optimizing Reservoir Computing for Reconstructing Ergodic Properties
Greg and Akira's paper