Reproduce “Context-Aware Learning to Rank with Self-Attention” 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
Przemysław Pobrotyn, Tomasz Bartczak, Mikołaj Synowiec, Radosław Białobrzeski, Jarosław Bojar
The context-aware ranker uses Transformer-style self-attention to let each result’s score depend on other results in the same list, rather than scoring items independently. On WEB30K, its ordinal-loss NDCG@5 was 53.00 ± 0.35, versus 48.84 ± 0.42 for a similarly sized MLP. On Allegro click logs, it improved relative NDCG@60 by 3.00%; this was an offline, not online, evaluation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.