Masked Diffusion Language Models are Strong and Steerable Text-Based World Models for Agentic RL

• URL: http://arxiv.org/abs/2607.16204v1
• Date: Thu, 07 May 2026 00:40:32 GMT
• Title: Masked Diffusion Language Models are Strong and Steerable Text-Based World Models for Agentic RL
• Abstract summary: We formalize text-based world modeling as a steerable transition-dynamics problem decomposed into initial state.<n>We curate 239,403 grounded state-action trajectories spanning nine open-source environments and twelve frontier model families.<n>We introduce a plug-and-play GRPO training framework with deterministic state checks, and perform zero-shot transfer ablations on three OOD environments.
• Score: 0.8546394944138487
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Recent growth in reinforcement learning (RL) has surfaced a need for diverse, specialized training environments. Hand-curated environments with fixed task and reward difficulties become ineffective signals as model performance improves, and sparse rewards over long horizons induce mode collapse on specific workflows or tool structures. World models that simulate environment states have matched pure rollout performance, making them promising for scaling diversity on-demand. However, autoregressive (AR) world models suffer from a left-to-right bias preventing conditioning on globally interdependent state anchors such as tool schemas, prior turns, and expected outcomes. We (i) formalize text-based world modeling as a steerable transition-dynamics problem decomposed into initial state, task context, tool schemas, domain rules, and steering directives, and (ii) curate 239,403 grounded state-action trajectories spanning nine open-source environments and twelve frontier model families. We compare AR LMs and masked diffusion language models (MDLMs), showing MDLMs, via bidirectional anchor-aware denoising, achieve better coherence, groundedness, and empirically validated rollout diversity than LLMs over 4x their parameter size, at comparable inference latency. We introduce a plug-and-play GRPO training framework with deterministic state checks, and perform zero-shot transfer ablations on three OOD environments (ScienceWorld, ALFWorld, AppWorld) across three 1.2B-7B agent backbones (LFM2.5, Qwen3, Mistral), achieving up to 47% absolute gains over baselines without environment-specific fine-tuning. We further conduct behavioral analysis of failure modes under adversarial scenarios and human evaluation on realism, outcome correctness, and training utility. We open-source our work to encourage research in this direction.

Related papers

• Generalization of World Models under Environmental Variability for Vision-based Quadrotor Navigation [10.364779390403337]
  World models, learned generative models that predict how an environment evolves, have become a promising tool for sample-efficient robot learning.<n>We conduct a systematic study using vision-based quadrotor navigation as a testbed problem, training DreamerV3-based world models under varying levels of environmental randomness.<n>We then deploy all world models and associated navigation policies on a real quadrotor in unseen environments, including an open-loop run where the model receives just 2.5s of real sensory input before all sensors are cut off.<n>Our results show that world model robustness during SSL pretraining is a strong predictor of sim-to-real
  arXiv  Detail & Related papers  (2026-06-03T15:38:36Z)
• LaST-R1: Reinforcing Robotic Manipulation via Adaptive Physical Latent Reasoning [90.86828952599147]
  We present LaST-R1, a novel reinforcement learning framework designed to harness "latent reasoning-before-acting" policies.<n>LaST-R1 achieves a near-perfect 99.9% average success rate on the LIBERO benchmark.<n>In real-world deployments, LaST-R1 yields up to a 22.5% average improvement over SOTA supervised fine-tuning approach.
  arXiv  Detail & Related papers  (2026-04-30T17:59:52Z)
• Reinforcement World Model Learning for LLM-based Agents [60.65003139516272]
  Reinforcement World Model Learning (RWML) is a self-conditioned method that learns action-supervised world models for LLM-based agents.<n>Our method aligns simulated next states produced by the model with realized next states observed from the environment.<n>We evaluate our method on ALFWorld and $2$ Bench and observe significant gains over the base model, despite being entirely self-supervised.
  arXiv  Detail & Related papers  (2026-02-05T16:30:08Z)
• Aligning Agentic World Models via Knowledgeable Experience Learning [68.85843641222186]
  We introduce WorldMind, a framework that constructs a symbolic World Knowledge Repository by synthesizing environmental feedback.<n>WorldMind achieves superior performance compared to baselines with remarkable cross-model and cross-environment transferability.
  arXiv  Detail & Related papers  (2026-01-19T17:33:31Z)
• Agent2World: Learning to Generate Symbolic World Models via Adaptive Multi-Agent Feedback [51.22403664895878]
  Agent2World is a tool-augmented multi-agent framework that achieves strong inference-time world-model generation.<n>It also serves as a data engine for supervised fine-tuning, by grounding generation in multi-agent feedback.
  arXiv  Detail & Related papers  (2025-12-26T18:54:14Z)
• From Word to World: Can Large Language Models be Implicit Text-based World Models? [82.47317196099907]
  Agentic reinforcement learning increasingly relies on experience-driven scaling.<n>World models offer a potential way to improve learning efficiency through simulated experience.<n>We study whether large language models can reliably serve this role and under what conditions they meaningfully benefit agents.
  arXiv  Detail & Related papers  (2025-12-21T17:28:42Z)
• Internalizing World Models via Self-Play Finetuning for Agentic RL [65.96875390986655]
  Large Language Models (LLMs) as agents often struggle in out-of-distribution (OOD) scenarios.<n>We show how to encode this world model by decomposing it into two components: state representation and transition modeling.<n>We introduce SPA, a simple reinforcement learning framework that cold-starts the policy via a Self-Play supervised finetuning stage to learn the world model.
  arXiv  Detail & Related papers  (2025-10-16T18:03:39Z)
• Reinforcement Learning for Machine Learning Model Deployment: Evaluating Multi-Armed Bandits in ML Ops Environments [0.0]
  We investigate whether reinforcement learning (RL)-based model management can manage deployment decisions more effectively.<n>Our approach enables more adaptive production environments by continuously evaluating deployed models and rolling back underperforming ones in real-time.<n>Our findings suggest that RL-based model management can improve automation, reduce reliance on manual interventions, and mitigate risks associated with post-deployment model failures.
  arXiv  Detail & Related papers  (2025-03-28T16:42:21Z)

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.