Connect the Dots: Training LLMs for Long-Lifecycle Agents with Cross-Domain Generalization Via Reinforcement Learning

• URL: http://arxiv.org/abs/2606.20002v1
• Date: Thu, 18 Jun 2026 09:38:45 GMT
• Title: Connect the Dots: Training LLMs for Long-Lifecycle Agents with Cross-Domain Generalization Via Reinforcement Learning
• Abstract summary: We present a framework for training large language models (LLMs) to "Connect the Dots" (CoD)<n>As an LLM-based AI agent gets deployed in an environment, it solves a long sequence of tasks while continuously exploring the environment.<n>We present proof-of-concept implementations of the CoD framework, including a GRPO-style RL algorithm with fine-grained credit assignment.
• Score: 82.3676770818744
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: This work presents a general framework for training large language models (LLMs) to "Connect the Dots" (CoD), a meta-capability required by long-lifecycle agents: as an LLM-based AI agent gets deployed in an environment, it solves a long sequence of tasks while continuously exploring the environment, learning from its own experiences, and iteratively self-updating its context about the environment, thereby achieving progressively better performance on future tasks conditioned on the updated context. Major components of the CoD framework include: (1) algorithm design and infrastructure for end-to-end reinforcement learning (RL) with long rollout sequences interleaving solve-task and update-context episodes; (2) tasks and environments for incentivizing and eliciting the targeted meta-capability in LLMs during training, as well as for faithfully measuring progress during evaluation. We present proof-of-concept implementations of the CoD framework, including a GRPO-style RL algorithm with fine-grained credit assignment, as well as tasks and environments tailored to the targeted meta-capability (rather than domain-specific LLM capabilities or standard task-by-task RL). Empirical results validate the efficacy of end-to-end RL training in the CoD setting, and demonstrate the potential for out-of-distribution generalization -- within the training domains, across different domains, and from CoD to Ralph-loop settings -- of the elicited meta-capability. Our investigation of CoD connects several lines of prior works, and opens up new opportunities for advancing LLMs and AI agents. To facilitate further research and applications, we release our implementations at \url{https://github.com/agentscope-ai/Trinity-RFT/tree/research/cod/examples/research_cod}.

Related papers

• A Practitioner's Guide to Multi-turn Agentic Reinforcement Learning [12.179148605060298]
  We study what actually works and what doesn't for training large language models as agents via multi-turn reinforcement learning.<n>We break down the design space into three inter-related pillars -- environment, reward, and policy.<n>We distill these findings into a training recipe that guides co-design across the three pillars.
  arXiv  Detail & Related papers  (2025-10-01T17:23:04Z)
• InternBootcamp Technical Report: Boosting LLM Reasoning with Verifiable Task Scaling [71.37579508777843]
  Large language models (LLMs) have revolutionized artificial intelligence by enabling complex reasoning capabilities.<n>To address this gap, we present InternBootcamp, an open-source framework comprising 1000+ domain-diverse task environments.
  arXiv  Detail & Related papers  (2025-08-12T05:00:00Z)
• Agentic Reinforced Policy Optimization [66.96989268893932]
  Large-scale reinforcement learning with verifiable rewards (RLVR) has demonstrated its effectiveness in harnessing the potential of large language models (LLMs) for single-turn reasoning tasks.<n>Current RL algorithms inadequately balance the models' intrinsic long-horizon reasoning capabilities and their proficiency in multi-turn tool interactions.<n>We propose Agentic Reinforced Policy Optimization (ARPO), a novel agentic RL algorithm tailored for training multi-turn LLM-based agents.
  arXiv  Detail & Related papers  (2025-07-26T07:53:11Z)
• Omni-Thinker: Scaling Multi-Task RL in LLMs with Hybrid Reward and Task Scheduling [66.0871543682453]
  We present Omni-Thinker, a unified reinforcement learning framework that scales large language models across diverse tasks.<n>Our scheduler orders tasks according to accuracy backward transfer (BWT), reducing forgetting and improving multi-task performance.
  arXiv  Detail & Related papers  (2025-07-20T01:50:16Z)
• UAS Visual Navigation in Large and Unseen Environments via a Meta Agent [0.13654846342364302]
  We propose a meta-curriculum training scheme to efficiently learn to navigate in large-scale urban environments.<n>We organize the training curriculum in a hierarchical manner such that the agent is guided from coarse to fine towards the target task.<n>In contrast to traditional reinforcement learning (RL), which focuses on acquiring a policy for a specific task, MRL aims to learn a policy with fast transfer ability to novel tasks.
  arXiv  Detail & Related papers  (2025-03-20T01:44:59Z)
• SWEET-RL: Training Multi-Turn LLM Agents on Collaborative Reasoning Tasks [110.20297293596005]
  Large language model (LLM) agents need to perform multi-turn interactions in real-world tasks.<n>Existing multi-turn RL algorithms for optimizing LLM agents fail to perform effective credit assignment over multiple turns while leveraging the generalization capabilities of LLMs.<n>We propose a novel RL algorithm, SWEET-RL, that uses a carefully designed optimization objective to train a critic model with access to additional training-time information.<n>Our experiments demonstrate that SWEET-RL achieves a 6% absolute improvement in success and win rates on ColBench compared to other state-of-the-art multi-turn RL algorithms.
  arXiv  Detail & Related papers  (2025-03-19T17:55:08Z)
• Words as Beacons: Guiding RL Agents with High-Level Language Prompts [6.7236795813629]
  Large Language Models (LLMs) as "teachers" guide the agent's learning process by decomposing complex tasks into subgoals. LLMs can provide subgoals to accomplish the task defined for the environment in a similar fashion to how a human would do. It is possible to query the LLM only during the training phase, enabling agents to operate within the environment without any LLM intervention.
  arXiv  Detail & Related papers  (2024-10-11T08:54:45Z)
• Discrete Factorial Representations as an Abstraction for Goal Conditioned Reinforcement Learning [99.38163119531745]
  We show that applying a discretizing bottleneck can improve performance in goal-conditioned RL setups. We experimentally prove the expected return on out-of-distribution goals, while still allowing for specifying goals with expressive structure.
  arXiv  Detail & Related papers  (2022-11-01T03:31:43Z)
• Continuous Coordination As a Realistic Scenario for Lifelong Learning [6.044372319762058]
  We introduce a multi-agent lifelong learning testbed that supports both zero-shot and few-shot settings. We evaluate several recent MARL methods, and benchmark state-of-the-art LLL algorithms in limited memory and computation. We empirically show that the agents trained in our setup are able to coordinate well with unseen agents, without any additional assumptions made by previous works.
  arXiv  Detail & Related papers  (2021-03-04T18:44:03Z)
• Meta Reinforcement Learning with Autonomous Inference of Subtask Dependencies [57.27944046925876]
  We propose and address a novel few-shot RL problem, where a task is characterized by a subtask graph. Instead of directly learning a meta-policy, we develop a Meta-learner with Subtask Graph Inference. Our experiment results on two grid-world domains and StarCraft II environments show that the proposed method is able to accurately infer the latent task parameter.
  arXiv  Detail & Related papers  (2020-01-01T17:34:00Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.