Evidence Receipt. Related Resources.
SCDP: Learning Humanoid Locomotion from Partial Observations via Mixed-Observation Distillation
Use This Via API or MCP
Use this Signal Canvas via API or MCP
Route this paper proof surface into REST, MCP, or developer workflows while preserving the same evidence receipt and related-resource context.
Page Freshness
Signal Canvas proof surface
Canonical route: /signal-canvas/scdp-learning-humanoid-locomotion-from-partial-observations-via-mixed-observation-distillation
- Proof freshness
- stale
- Proof status
- unverified
- Display score
- 8/10
- Last proof check
- 2026-03-19
- Score updated
- 2026-04-02
- Score fresh until
- 2026-05-02
- References
- 0
- Source count
- 0
- Coverage
- 33%
This page is showing the last landed evidence receipt and score bundle because the latest proof data is outside the freshness window.
Agent Handoff
SCDP: Learning Humanoid Locomotion from Partial Observations via Mixed-Observation Distillation
Canonical ID scdp-learning-humanoid-locomotion-from-partial-observations-via-mixed-observation-distillation | Route /signal-canvas/scdp-learning-humanoid-locomotion-from-partial-observations-via-mixed-observation-distillation
REST example
curl https://sciencetostartup.com/api/v1/agent-handoff/signal-canvas/scdp-learning-humanoid-locomotion-from-partial-observations-via-mixed-observation-distillationMCP example
{
"tool": "search_signal_canvas",
"arguments": {
"mode": "paper",
"paper_ref": "scdp-learning-humanoid-locomotion-from-partial-observations-via-mixed-observation-distillation",
"query_text": "Summarize SCDP: Learning Humanoid Locomotion from Partial Observations via Mixed-Observation Distillation"
}
}source_context
{
"surface": "signal_canvas",
"mode": "paper",
"query": "SCDP: Learning Humanoid Locomotion from Partial Observations via Mixed-Observation Distillation",
"normalized_query": "2603.09574",
"route": "/signal-canvas/scdp-learning-humanoid-locomotion-from-partial-observations-via-mixed-observation-distillation",
"paper_ref": "scdp-learning-humanoid-locomotion-from-partial-observations-via-mixed-observation-distillation",
"topic_slug": null,
"benchmark_ref": null,
"dataset_ref": null
}Preparing verified analysis
Dimensions overall score 8.0
GitHub Code Pulse
No public code linked for this paper yet.
Claim map
- Evidencepartial
SCDP decouples sensing from supervision through mixed-observation training: diffusion model conditions on sensor histories while being supervised to predict privileged future state-action trajectories, enforcing the model to infer the motion dynamics under partial observability.
ImplicationpartialDirectly and explicitly stated in the abstract as a core contribution of the method.
Verificationpartialpartial
- Evidencepartial
In simulation, SCDP achieves near-perfect success on velocity control (99-100%)
ImplicationpartialExplicitly stated numeric result for a specific task in the abstract.
Verificationpartialpartial
- Evidencepartial
and 93% tracking success in AMASS test set
ImplicationpartialExplicitly stated numeric result for a specific task and dataset in the abstract.
Verificationpartialpartial
- Evidencepartial
performing comparable to privileged baselines while using only onboard sensors.
ImplicationpartialDirectly stated comparative result in the abstract, though 'comparable' is qualitative.
Verificationpartialpartial
- Evidencepartial
Finally, we deploy the trained policy on a real G1 humanoid at 50 Hz
ImplicationpartialDirectly stated deployment result with specific robot model and frequency.
Verificationpartialpartial
- Evidencepartial
demonstrating robust real robot locomotion without external sensing or state estimation.
ImplicationpartialDirectly stated claim about real-world performance, though 'robust' is a qualitative assessment.
Verificationpartialpartial
- Evidencepartial
enforcing the model to infer the motion dynamics under partial observability.
ImplicationpartialDirectly stated technical mechanism and its purpose in the method description.
Verificationpartialpartial
- Evidencepartial
We further develop restricted denoising, context distribution alignment, and context-aware attention masking to encourage implicit state estimation within the model and to prevent train-deploy mismatch.
ImplicationpartialDirectly stated technical components and their stated purpose in the abstract.
Verificationpartialpartial
