Reproduce “Enhanced Recurrent Neural Tangent Kernels for Non-Time-Series Data” 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

Sina Alemohammad, Randall Balestriero, Zichao Wang, Richard Baraniuk

The authors extend infinite-width recurrent neural network kernels to classify fixed-length non-time-series feature vectors, deriving bidirectional and average-pooled variants and evaluating kernel SVMs on 90 UCI datasets. RNN-P, tested in default and reversed feature order, achieves mean test accuracy of 82.34%, versus 82.25% for H-γ-exp. Its P95 is lower than H-γ-exp and BI-RNN-AVG, so it does not lead every aggregate measure.

Summary by alphaXiv.

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