Reproduce “Approximating exponential family models (not single distributions) with a two-network architecture” 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

Sean R. Bittner, John P. Cunningham

The authors introduce an exponential-family network that maps natural parameters to flow settings, producing approximate distributions across a family rather than optimizing each dataset separately. In neural spike data, three example latent-intensity trajectories look similar to separately fitted flows. The comparison is limited to a specified log-Gaussian Poisson model; amortization benefits depend on workload and target quality.

Summary by alphaXiv.

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