Reproduce “Hierarchical Text Classification Using Contrastive Learning Informed Path Guided Hierarchy” 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
Neeraj Agrawal, Saurabh Kumar, Priyanka Bhatt, Tanishka Agarwal
HTC-CLIP combines hierarchy-aware text representations with a path-guided classifier. It trains BERT using hierarchy-selected positive text views and contrastive learning, alongside linear and hierarchical classification heads. At inference, it takes the maximum of their per-class probabilities. On WOS test Macro-F1, it reports 81.64 versus 80.30 for its HGCLR reimplementation; evidence covers two public benchmarks, not deployment.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.