RoboTALES: Learning Reasoning-Guided Robot Policies via Task-Aligned Simulated Futures

  • 类型:arxiv
  • 标识:2607.06018
  • 链接:https://arxiv.org/abs/2607.06018
  • 主分类:multimodal
  • 形态:method
  • 被引:0
  • 被引来源:Semantic Scholar + OpenAlex
  • S2被引:0
  • OpenAlex被引:0
  • 影响力被引:0
  • TLDR:This work proposes RoboTALES, a single-stage framework that learns task-aligned simulated futures and uses them to train robot policies and introduces two key innovations: a hierarchical LLM-based planner that breaks complex tasks into a sequence of subgoals to guide the model's imagination and a VLM-based critic that evaluates these ``imagined'' futures.
  • OpenAlex ID:W7167708930
  • OpenAlex DOI:10.48550/arxiv.2607.06018
  • DOI:10.48550/arxiv.2607.06018
  • DOI来源:OpenAlex
  • 开放获取:green
  • 开放获取链接:https://doi.org/10.48550/arxiv.2607.06018
  • OpenAlex更新:2026-07-19
  • 待LLM分类:否
  • 标题中文:RoboTALES:通过任务对齐的模拟未来学习推理引导的机器人策略
  • TLDR中文:提出 RoboTALES,一个学习任务对齐模拟未来并据此训练机器人策略的单阶段框架,引入两项关键创新:基于 LLM 的分层规划器,将复杂任务拆解为子目标序列以引导模型的"想象";基于 VLM 的评判器,用于评估这些"想象"出的未来。
  • 来源文件
  • /inbox/tom/_candidates/2026-07-10-agent-rag-longcontext-candidates.json
  • [S2 enrich]
  • [OpenAlex backfill]