TeachAnything: A Multimodal Crowdsourcing Platform for Training Embodied AI Agents in Symmetrical Reality

TL;DR

TeachAnything is a multimodal crowdsourcing platform that collects diverse demonstration data in physics-simulated environments to train embodied AI agents for Symmetrical Reality.

Contribution

It introduces a three-stage demonstration paradigm and a cloud-based platform integrating virtual and physical interactions for embodied AI training.

Findings

Developed a scalable platform capable of collecting diverse multimodal demonstrations.

Unified virtual and physical interactions through physics simulation and methodological design.

Provides a practical foundation for training embodied agents in Symmetrical Reality.

Abstract

Symmetrical Reality (SR) is emerging as a future trend for human-agent coexistence, placing higher demands on agents to acquire human-like intelligence. It calls for richer and more diverse human guidance. We introduce a three-stage demonstration paradigm integrating multimodal demonstration signals. Building on this paradigm, we developed TeachAnything, a cloud-based, crowdsourcing-oriented demonstration platform with physics simulation capable of collecting diverse demonstration data across varied scenes, tasks, and embodiments. By unifying virtual and physical interactions through both methodological design and physics simulation, the system serves as a practical foundation for developing embodied agents aligned with Symmetrical Reality.