paper-with-me

홈 › Papers

InftyThink+: Effective and Efficient Infinite-Horizon Reasoning via Reinforcement Learning

2026-02-06 · Yuchen Yan, Liang Jiang, Jin Jiang, Shuaicheng Li, Zujie Wen, Zhiqiang Zhang, Jun Zhou, Jian Shao, Yueting Zhuang, Yongliang Shen arxiv

Large reasoning models achieve strong performance by scaling inference-time chain-of-thought, but this paradigm suffers from quadratic cost, context length limits, and degraded reasoning due to lost-in-the-middle effects. Iterative reasoning mitigates these issues by periodically summarizing intermediate thoughts, yet existing methods rely on supervised learning or fixed heuristics and fail to optimize when to summarize, what to preserve, and how to resume reasoning. We propose InftyThink+, an end-to-end reinforcement learning framework that optimizes the entire iterative reasoning trajectory, building on model-controlled iteration boundaries and explicit summarization. InftyThink+ adopts a two-stage training scheme with supervised cold-start followed by trajectory-level reinforcement learning, enabling the model to learn strategic summarization and continuation decisions. Experiments on DeepSeek-R1-Distill-Qwen-1.5B show that InftyThink+ improves accuracy by 21% on AIME24 and outperforms conventional long chain-of-thought reinforcement learning by a clear margin, while also generalizing better to out-of-distribution benchmarks. Moreover, InftyThink+ significantly reduces inference latency and accelerates reinforcement learning training, demonstrating improved reasoning efficiency alongside stronger performance.

📄 PDF Abstract BibTeX arXiv:2602.06960

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement Learning

Similar Papers 제목 키워드 기반

Directional Attractors in LLM Reasoning: How Similarity Retrieval Steers Iterative Summarization Based Reasoning

2025-12-22 · Cagatay Tekin, Charbel Barakat, Luis Joseph Luna Limgenco arxiv

Iterative summarization based reasoning frameworks such as InftyThink enable long-horizon reasoning in large language models (LLMs) by controlling context growth, but they repeatedly regenerate similar reasoning strategi…

InftyThink: Breaking the Length Limits of Long-Context Reasoning in Large Language Models

2025-03-09 · Yuchen Yan, Yongliang Shen, Yang Liu, Jin Jiang 외

Advanced reasoning in large language models has achieved remarkable performance on challenging tasks, but the prevailing long-context reasoning paradigm faces critical limitations: quadratic computational scaling with se…

Computational EfficiencyMath

Beyond dynamic programming

2023-06-26 · Abhinav Muraleedharan

In this paper, we present Score-life programming, a novel theoretical approach for solving reinforcement learning problems. In contrast with classical dynamic programming-based methods, our method can search over non-sta…

reinforcement-learningReinforcement Learning

Offline Reinforcement Learning with Universal Horizon Models

2026-05-15 · Hojun Chung, Junseo Lee, Songhwai Oh arxiv

Model-based reinforcement learning (RL) offers a compelling approach to offline RL by enabling value learning on imagined on-policy trajectories. However, it often suffers from compounding errors due to repeated model in…

Reinforcement LearningOffline RL

Infinite Time Horizon Safety of Bayesian Neural Networks

2021-11-04 · NeurIPS 2021 12 · Mathias Lechner, Đorđe Žikelić, Krishnendu Chatterjee, Thomas A. Henzinger

Bayesian neural networks (BNNs) place distributions over the weights of a neural network to model uncertainty in the data and the network's prediction. We consider the problem of verifying safety when running a Bayesian …

reinforcement-learningReinforcement LearningReinforcement Learning (RL)Safe Exploration