method · first seen 12 Oct 2026
contextual dueling bandit
A bandit setting where the learner chooses between pairs of options in a context using only pairwise preference feedback.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Online LLM Personalization from Pairwise Preferences Across UsersPapers · 12 Oct 2026 · Romain Puech, Georgy Salakhutdinov, Dimitris Bertsimas
Latest sources
- Online LLM Personalization from Pairwise Preferences Across UsersPapers · 12 Oct 2026 · Romain Puech, Georgy Salakhutdinov, Dimitris Bertsimas