TSBPO: Tree-Structured Branching for Policy Optimization in Interactive Language Agents
cs.AI, cs.CL, cs.LG
Submitted: 2026-10-30
Updated: 2026-11-05
Comments: v2: corrected experiment configuration description, revised abstract and claims; results unchanged
Project page: https://wouterkool.github.io/publication/buy-4-samples-freebaseline
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- ALFWorld: Aligning Text and Embodied Environments for Interactive Learning
- Proximal Policy Optimization Algorithms
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- TreePO: Bridging the Gap of Policy Optimization and Efficacy and Inference Efficiency with Heuristic Tree-based Modeling
- Understanding R1-Zero-Like Training: A Critical Perspective
- BiPACE: Bisimulation-Guided Policy Optimization with Action Counterfactual Estimation for LLM Agents
- Dynamics-Predictive Sampling for Active RL Finetuning of Large Reasoning Models
- DAPO: An Open-Source LLM Reinforcement Learning System at Scale
- RAGEN: Understanding Self-Evolution in LLM Agents via Multi-Turn Reinforcement Learning
- WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents
- ReAct: Synergizing Reasoning and Acting in Language Models
- Reflexion: Language Agents with Verbal Reinforcement Learning
- Qwen2.5 Technical Report
- Trust Region Policy Optimization
- High-Dimensional Continuous Control Using Generalized Advantage Estimation
- ArCHer: Training Language Model Agents via Hierarchical Multi-Turn RL
- Reinforcement Learning for Long-Horizon Interactive LLM Agents
- RUDDER: Return Decomposition for Delayed Rewards
- Hindsight Credit Assignment
- Hierarchy-of-Groups Policy Optimization for Long-Horizon Agentic Tasks
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents