Reproduce “Memory flows: geometry and dynamics of sequential retrieval in input-driven Hopfield networks” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Simone Betteti, Matteo Martin, Morten Gram Pedersen, Sandro Zampieri, Giacomo Baggio

The authors develop a Hopfield memory using slow feedback to move retrieval along a graph of allowed transitions. The retrieved pattern loses salience while its successor gains it, allowing the current memory to lose stability. For orthogonal memories, the SmoothF two-memory analysis finds fixed points above a critical gain of about 12.74, but sustained retrieval also requires sufficient initial salience.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.