Large-Scale Egocentric Multimodal Dataset for Embodied AI

Xperience-10M is a large-scale egocentric multimodal dataset of human experience created by ropedia-ai. It is designed for research in embodied AI, robotics, and world models. The dataset was last updated on March 20, 2026.

Use Cases

• Training world models for robotics using egocentric video and motion capture (mocap) data.
• Developing multimodal agents that process first-person video, 3D data, and language (en) for interactive intelligence tasks.
• Researching human motion and spatial reasoning from point cloud and 4D data in an egocentric context.
• Benchmarking video classification or depth estimation models on first-person perspective footage.

Strengths

• Dataset scale is indicated as between 1 million and 10 million samples based on platform size tags.
• Covers multiple modalities including video, 3D, point cloud, mocap, and language for a holistic representation of experience.

Limitations

• Specific row count, column details, and file size are unknown, hindering precise resource planning.
• Geographic coverage may be biased, as a platform tag indicates a region of 'us'.

Provenance

Source

ropedia-ai

Collection Method

Assumed to be collected from egocentric sensors capturing human experience; specific methodology is not detailed.

Freshness

Last updated on March 20, 2026, indicating recent maintenance.

Geography

Platform tags suggest a primary region of 'us' (United States).