Stage alignment
Each stage-level environment uses the matching physical success predicate from the long task.
Aspirate from a sample cup, return the pipette, then dispense into a target well and return it again.
Stage 01 / 02DualSharpaPipetteTransfer-v1Canonical instructions, handedness, action width, horizon, and success conditions are drawn from the current task registry and success-contract documentation.
| # | Stage / environment | Hand | Action | Max steps | Success summary |
|---|---|---|---|---|---|
| 01 | Aspirate liquidSharpaAngledMobilePipetteAspirate-v1 | Right | 29 | 800 | Prime the plunger, dip the tip, release the plunger, then return and release the pipette. |
| 02 | Dispense liquidSharpaAngledMobilePipetteDispense-v1 | Left | 29 | 1100 | Insert the tip within 12 mm of the tube axis, depress the plunger, then return and release the pipette. |
Stage success in the training scene (%, 50 episodes). Best scores in each row are highlighted.
| Stage | PPO | RLPD | ACT | Diffusion Policy | GR00T | π0.5 |
|---|---|---|---|---|---|---|
| AspirateLiquid | 0.0 | 0.0 | 40.0 | 2.0 | 8.0 | 26.0 |
| DispenseLiquid | 0.0 | 0.0 | 0.0 | 2.0 | 8.0 | 0.0 |
| Stage | PPO | RLPD | ACT | Diffusion Policy | GR00T | π0.5 |
|---|---|---|---|---|---|---|
| AspirateLiquid | 0.0 ± 0.0 | 0.0 ± 0.0 | 8.7 ± 8.1 | 1.3 ± 1.2 | 2.0 ± 2.0 | 34.0 ± 6.9 |
| DispenseLiquid | 0.0 ± 0.0 | 0.0 ± 0.0 | 0.0 ± 0.0 | 0.0 ± 0.0 | 2.0 ± 2.0 | 4.0 ± 0.0 |
| Task | PPO | RLPD | ACT | Diffusion Policy | GR00T | π0.5 |
|---|---|---|---|---|---|---|
| Pipette transfer | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
Source: manuscript experiment tables. Compare all methods and tasks →
Each stage-level environment uses the matching physical success predicate from the long task.
Hold counters or latches prevent unstable contact from being counted as task completion.
State and action traces can be restored and replayed for validation and offline rendering.