Reproduce “Canonical Correlation Neural 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
Andrea Cavallo, Sundeep Prabhakar Chepuri, Charilaos Kanatsoulis, Elvin Isufi
CANE uses separate covariance-based stages to process each view and their cross-view relationships before a task-trained readout predicts. On MHEALTH action classification, it reports 95.3±1.1% accuracy, versus 94.2±0.4% for a two-view VNN and TOCCA. It is best on six of seven tasks; on ABIDE age prediction, it is within one standard deviation of the best model.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.