Research map / Reinforcement learning / Policy gradient and actor-critic
Batch Policy Gradient Methods for Improving Neural Conversation Models
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 |
|---|---|---|---|
| Initialize the bot with maximum-likelihood training (ablation) | Batch Policy Gradient · Synthetic sequence reproduction | convergence | Helped |
| Use importance sampling for off-policy updates (ablation) | Batch Policy Gradient · Synthetic sequence reproduction | learning performance | Helped |
| Explicitly account for future rewards (ablation) | Batch Policy Gradient · Synthetic sequence reproduction | convergence speed | Helped |
| Change the value-function estimator (ablation) | Batch Policy Gradient · Synthetic sequence reproduction with less data | learning performance | No clear effect |
| Use lambda < 1 (ablation) | Batch Policy Gradient · Synthetic sequence reproduction with noisy rewards | learning performance | Helped |
| Apply Batch Policy Gradient after maximum-likelihood training | Restaurant recommendation chat-bots · Restaurant recommendations | human-rated response quality | Helped |
Closest papers by meaning
- Natural Language Actor-Critic: Scalable Off-Policy Learning in Language Space (NeurIPS 2026 · similarity 0.86)
- Dialogue for Prompting: A Policy-Gradient-Based Discrete Prompt Generation for Few-Shot Learning (AAAI 2024 · similarity 0.86)
- Offline RL by Reward-Weighted Fine-Tuning for Conversation Optimization (NeurIPS 2025 · similarity 0.86)
- MA-RLHF: Reinforcement Learning from Human Feedback with Macro Actions (ICLR 2025 · similarity 0.86)
- Asymmetric REINFORCE for off-Policy Reinforcement Learning: Balancing positive and negative rewards (NeurIPS 2025 · similarity 0.85)
- Simulation-Free Hierarchical Latent Policy Planning for Proactive Dialogues (AAAI 2025 · 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.