Reproduce “CADET: Context-Conditioned Ads CTR Prediction With a Decoder-Only Transformer” 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
David Pardoe, Neil Daftary, Miro Furtado, Aditya Aiyer, Yu Wang, Liuqing Li, Tao Song, Lars Hertel and 15 more
CADET estimates ad clicks while accounting for feed position, which is determined after initial scoring. A shared decoder-only sequence model produces scores for two position buckets; the auction uses the score matching the placement. In a LinkedIn homefeed sponsored-updates A/B test, CADET achieved an 11.04% relative CTR lift over LiRank, evaluating the system rather than its context heads alone.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.