Research map / Reinforcement learning / Offline RL
A General Offline Reinforcement Learning Framework for Interactive Recommendation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Apply dual constraints to SDAC to reduce policy mismatch across future steps | SDAC · RecSys | HR@5: .2341 -> .3272 | Helped |
| Apply dual constraints to SDAC to reduce policy mismatch across future steps | SDAC · Kaggle | HR@5: .1669 -> .2659 | Helped |
| Add dual constraints rather than use supervised learning alone | Interactive recommendation · RecSys | NDCG@10: .2279 -> .2593 | Helped |
| Add dual constraints rather than use supervised learning alone | Interactive recommendation · Kaggle | NDCG@10: .1878 -> .2351 | Helped |
| Use off-policy DQN instead of supervised learning | Interactive recommendation · RecSys | HR@5: .2876 -> .2134 | Hurt |
| Use off-policy policy gradient instead of supervised learning | Interactive recommendation · Kaggle | HR@5: .2233 -> .1585 | Hurt |
| Use stochastic discrete policy learning rather than DQN | Offline recommendation · RecSys | HR@5: .2134 -> .2341 | Helped |
| Add supervised regularization to SDAC | SDAC · RecSys | HR@5: .2341 -> .3197 | Helped |
| Add support constraints to SDAC | SDAC · RecSys | HR@5: .2341 -> .2987 | Helped |
| Add policy constraints to SDAC | SDAC · Kaggle | HR@5: .1669 -> .2354 | Helped |
| Use reward extrapolation rather than supervised regularization | Offline recommendation with sparse purchase feedback · RecSys purchase | HR@5: .4341 -> .4372 | Helped |
| Adapt the supervised-regularization multiplier using dual gradient descent | SR · Two datasets | performance | Helped |
Closest papers by meaning
- Text-Based Interactive Recommendation via Offline Reinforcement Learning (AAAI 2022 · similarity 0.95)
- A Model-Based Reinforcement Learning with Adversarial Training for Online Recommendation (NeurIPS 2019 · similarity 0.90)
- On the Unexpected Effectiveness of Reinforcement Learning for Sequential Recommendation (ICML 2024 · similarity 0.88)
- MTRec: Learning to Align with User Preferences via Mental Reward Models (NeurIPS 2025 · similarity 0.87)
- Influential Exemplar Replay for Incremental Learning in Recommender Systems (AAAI 2024 · similarity 0.87)
- Objective Shaping with Hard Negatives: Windowed Partial AUC Optimization for RL-based LLM Recommenders (NeurIPS 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.