WCM: World-Cognition Model for Generalizable Human-Robot Interaction

2026-07-24

AI Summary

This work addresses the limited interpretability and lack of user intervention capabilities in existing robot control methods, which hinder natural and teachable human–robot interaction. The authors propose an embodied agent based on a modular SLAK (Sensing–Logic–Action–Knowledge) architecture with an asynchronous runtime, enabling concurrent reasoning, dialogue, and execution. A human–robot co-teaching framework coupled with a chain-of-thought supervision mechanism integrates interactive teaching episodes with autonomously collected execution data to continuously refine the model and enhance task generalization. Evaluated across nine real-world tasks—including those not fine-tuned and long-horizon tasks acquired through teaching—the system achieves an average success rate of 73.8%.

Abstract

Language agents can now interact fluently with users in software, but robots still struggle to bring comparable interaction to physical tasks. Current robot-control paradigms, including vision-language-action policies and world-model-based planners, are mainly optimized for instruction execution, leaving users with little visibility into why an action is chosen and few mechanisms to redirect, correct, or teach the robot through interaction. To solve this problem, we present the World-Cognition Model (WCM), a human-centered embodied agent built on the SLAK architecture (Sensing, Logic, Action, and Knowledge) and an asynchronous runtime. SLAK separates perception, reasoning, control, and memory, while the runtime allows reasoning, dialogue, and execution to proceed concurrently. WCM further introduces a human-in-the-loop teaching mode that enables users to interactively teach the robot difficult or long-horizon tasks. Teaching episodes and autonomous task rollouts are refined into chain-of-thought supervision to continually improve the model. WCM achieves a 73.8% average success rate across nine real-world human-robot interaction tasks, including tasks held out from CoT fine-tuning and a long-horizon task learned through teaching.

Problem

• human-robot interaction
• world model
• explainability
• teachability
• embodied AI

Innovation

• World-Cognition Model
• SLAK architecture
• human-in-the-loop teaching
• chain-of-thought supervision
• asynchronous runtime