Research map / Reinforcement learning / Policy gradient and actor-critic
Large-Scale Interactive Recommendation with Tree-Structured Policy Gradient
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 |
|---|---|---|---|
| Use the tree-structured policy gradient framework instead of the strongest reported RL baseline | Interactive recommendation · MovieLens, α=0.1 | Reward: 0.2379 -> 0.2472 | Helped |
| Use the tree-structured policy gradient framework instead of the strongest reported RL baseline | Interactive recommendation · Netflix, α=0.1 | Reward: 0.2044 -> 0.2544 | Helped |
| Use TPGR rather than DQN-R for action selection | Netflix recommendation · Netflix, α=0.0 | seconds per 10^6 decisions: 34.6 -> 3.9 | Helped |
| Use TPGR rather than DDPG-KNN with k=0.1N | Interactive recommendation · MovieLens | seconds per 10^6 decisions: 200.4 -> 3.4 | Helped |
| Use PCA-based clustering with rating-based item representations rather than other tested clustering combinations (ablation) | TPGR · Netflix, α=0.1 | Reward | Helped |
| Set tree depth to 2 rather than the other tested depths (ablation) | TPGR · Netflix | Reward | Helped |
| Use a tree structure rather than a single policy network (depth 1) (ablation) | TPGR · Netflix | training time | Helped |
| Increase the reward trade-off factor α (ablation) | TPGR · MovieLens and Netflix | Reward | Helped |
Closest papers by meaning
- The Adaptive Q-Network for Recommendation Tasks with Dynamic Item Space (AAAI 2025 · similarity 0.87)
- Contrastive Representation for Interactive Recommendation (AAAI 2025 · similarity 0.87)
- Objective Shaping with Hard Negatives: Windowed Partial AUC Optimization for RL-based LLM Recommenders (NeurIPS 2026 · similarity 0.87)
- On the Unexpected Effectiveness of Reinforcement Learning for Sequential Recommendation (ICML 2024 · similarity 0.86)
- ConvRec-R1: Training LLM-based Conversational Recommender Systems with Reinforcement Learning (ICLR 2026 · similarity 0.86)
- UNEX-RL: Reinforcing Long-Term Rewards in Multi-Stage Recommender Systems with UNidirectional EXecution (AAAI 2024 · similarity 0.85)
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.