Embodied AI

Embodied AI datasets are trajectories, not frame collections

Evaluate robot and simulated-agent datasets through embodiment, sensors, actions, task context, outcomes, coverage, and rights.

Synthetic schema sampleObservation → action → next state
01
00.000sObservergb.player_view
02
00.050sActmove.forward + camera.delta
03
00.100sTransitionpose.x +0.84 m
04
00.150sOutcomeprogress 0.31 → 0.36

The values are synthetic and illustrate the public specification. They are not measurements from a commercial dataset.

Direct answer

An embodied AI dataset records an agent interacting with an environment through a specific body and control interface. The useful unit is a trajectory connecting observations, language or goals, actions, state changes, and outcomes.

Embodiment changes the meaning of every action

The same high-level instruction can require different kinematics, sensors, control rates, and safety constraints across a single arm, mobile manipulator, bimanual platform, quadruped, or simulated agent.

A release should identify morphology, joints, end effectors, camera placement, calibration, controller, coordinate frames, payload assumptions, and software or firmware version.

Demonstrations need context and negative evidence

Successful demonstrations teach one route through a task. Failures, recoveries, corrections, aborted attempts, and human interventions reveal decision boundaries and operational limits.

Episode records should preserve who or what controlled the system, how instructions were delivered, whether assistance occurred, why the episode ended, and which success test was applied.

  • Teleoperation, autonomous, scripted, or mixed control
  • Task instruction, scene, object, and initial-state metadata
  • Raw sensors, calibrated observations, and validity masks
  • Success, failure, correction, intervention, and safety events

A unified format should not erase source differences

Cross-dataset loaders make large mixtures practical. They should not discard original action semantics, sampling rates, licenses, collection methods, or quality decisions.

Normalize through explicit adapters and retain source-level provenance. Buyers and researchers need to know which transformations enable comparison and which differences remain material.

Questions this page should settle

What is an embodied AI dataset?

It is a collection of interaction trajectories grounded in an agent's body, sensors, action interface, environment, tasks, and outcomes.

Why does robot metadata matter?

Actions only have meaning relative to morphology, coordinate frames, units, control rate, calibration, and controller behavior.

Can datasets from different robots be combined?

Yes, but each source needs explicit adapters, embodiment metadata, provenance, compatible rights, and evaluation that tests transfer across robots rather than hiding differences.

Bring us the training target.

We will map it to the observations, actions, state, outcomes, rights, and delivery evidence the dataset must contain.

Request access