Scholarlink

<p>The realization of intelligent robots, operating autonomously and interacting with other intelligent agents, human or artificial, requires the integration of environment perception, reasoning, and action. Classic Artificial Intelligence techniques for this purpose, focusing on <i>symbolic</i> approaches, have long-ago hit the scalability wall on compute and memory costs. Advances in Large Language Models in the past decade (<i>neural</i> approaches) have resulted in unprecedented displays of capability, at the cost of control, explainability, and interpretability. Large Action Models aim at extending Large Language Models to encompass the full perception, reasoning, and action cycle; however, they typically require substantially more comprehensive training and suffer from the same deficiencies in reliability. Here, we demonstrate a prototype Large Action Model built by composing off-the-shelf foundation models, where control, interpretability, and explainability are improved by incorporating <i>symbolic wrappers</i> (symbolic validation layers) and associated verification on model outputs, resulting in a verifiable <i>neuro-symbolic</i> (hybrid neural–symbolic) solution for intelligent robots. Our experiments on a multi-modal robot demonstrate that Large Action Model intelligence does not require massive end-to-end training, but can be achieved by integrating efficient perception models with a logic-driven core. We find that driving action execution through the generation of Planning Domain Definition Language (PDDL) code enables a human-in-the-loop verification stage that effectively mitigates action hallucinations. These results can support practitioners in the design and development of robotic Large Action Models across novel industries, and shed light on the ongoing challenges that must be addressed to ensure safety in the field.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Symbolic wrappers enable precise human in the loop operation for large action models

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月7日

查看全文
  • Kanisorn Sangchai,
  • Methasit Boonpun,
  • Withawin Kraipetchara,
  • Paulo Garcia

摘要

The realization of intelligent robots, operating autonomously and interacting with other intelligent agents, human or artificial, requires the integration of environment perception, reasoning, and action. Classic Artificial Intelligence techniques for this purpose, focusing on symbolic approaches, have long-ago hit the scalability wall on compute and memory costs. Advances in Large Language Models in the past decade (neural approaches) have resulted in unprecedented displays of capability, at the cost of control, explainability, and interpretability. Large Action Models aim at extending Large Language Models to encompass the full perception, reasoning, and action cycle; however, they typically require substantially more comprehensive training and suffer from the same deficiencies in reliability. Here, we demonstrate a prototype Large Action Model built by composing off-the-shelf foundation models, where control, interpretability, and explainability are improved by incorporating symbolic wrappers (symbolic validation layers) and associated verification on model outputs, resulting in a verifiable neuro-symbolic (hybrid neural–symbolic) solution for intelligent robots. Our experiments on a multi-modal robot demonstrate that Large Action Model intelligence does not require massive end-to-end training, but can be achieved by integrating efficient perception models with a logic-driven core. We find that driving action execution through the generation of Planning Domain Definition Language (PDDL) code enables a human-in-the-loop verification stage that effectively mitigates action hallucinations. These results can support practitioners in the design and development of robotic Large Action Models across novel industries, and shed light on the ongoing challenges that must be addressed to ensure safety in the field.