8 papers
Credit-assigned Policy Gradient for Early Stage Retrieval in Two-stage Ranking
Haruka Kiyohara, Mihaela Curmei, Ariel Evnine +7
Large-scale search, recommendation, and retrieval-augmented generation (RAG) systems typically employ a two-stage architecture: an early-stage ranker (ESR) generates a candidate se…
Safely Exploring Novel Actions in Recommender Systems via Deployment-Efficient Policy Learning
Haruka Kiyohara, Yusuke Narita, Yuta Saito +2
In many real recommender systems, novel items are added frequently over time. The importance of sufficiently presenting novel actions has widely been acknowledged for improving lon…
Policy Design for Two-sided Platforms with Participation Dynamics
Haruka Kiyohara, Fan Yao, Sarah Dean
In two-sided platforms (e.g., video streaming or e-commerce), viewers and providers engage in interactive dynamics: viewers benefit from increases in provider populations, while pr…
Prompt Optimization with Logged Bandit Data
Haruka Kiyohara, Daniel Yiming Cao, Yuta Saito +1
We study how to use naturally available user feedback, such as clicks, to optimize large language model (LLM) pipelines for generating personalized sentences using prompts. Naive a…
Effective Off-Policy Evaluation and Learning in Contextual Combinatorial Bandits
Tatsuhiro Shimizu, Koichi Tanaka, Ren Kishimoto +3
We explore off-policy evaluation and learning (OPE/L) in contextual combinatorial bandits (CCB), where a policy selects a subset in the action space. For example, it might choose a…
Towards Assessing and Benchmarking Risk-Return Tradeoff of Off-Policy Evaluation
Haruka Kiyohara, Ren Kishimoto, Kosuke Kawakami +3
Off-Policy Evaluation (OPE) aims to assess the effectiveness of counterfactual policies using only offline logged data and is often used to identify the top-k promising policies fo…