Reproduce “An associative memory for the on-line recognition and prediction of temporal sequences” 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
J. Bose, S.B. Furber, J.L. Shapiro
The authors propose a sequence-memory model combining a shift register of recent symbols with a neural layer that transforms and carries older context. It uses the current symbol and updated context to retrieve a prediction. After learning and a second presentation, it recovered a mean of 22.03 of 23 symbols and perfect recovery in 10 of 30 runs, exceeding the variants in this simulation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.