Reproduce “HIERA: Hierarchical Multi-Agent Relevance Assessment for Content Discovery Systems” 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
Pritom Saha Akash, Phanideep Gampa, Chao Shen, Ying Chen, Sheikh Muhammad Sarwar
HIERA assesses search relevance by coordinating a judge with query, item, and relation analyzers. The judge requests focused analyses, which the relation analyzer synthesizes before the final label. Across five entertainment and product-search datasets, HIERA outperforms 11 baselines; on EVS, accuracy is 0.713 versus 0.517 for the strongest baseline. An ablation supports a benefit from this coordination in tested systems.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.