Reproduce “BIRDNet: Mining and Encoding Boolean Implication Knowledge Graphs as Interpretable Deep 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
Tirtharaj Dash
BIRDNet classifies samples using relationships mined from gene and protein measurements to structure its hidden layers. Each hidden unit represents a relationship between two features and is masked to connect only to them; a dense classifier head remains. Across six benchmarks, AUROC stays within 0.02 of the strongest dense baseline, but controls show mined pairings add no predictive advantage.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.