[yuecideng]

Description

This PR updates docs/source/resources/roadmap.md to reflect current development priorities across several product areas.

Changes:

• Rendering: Add an in-progress item for optimizing batch rendering throughput and latency on graphics cards (e.g. RTX series) and compute cards (e.g. A series).
• Physics: Add an in-progress item for Newton backend support, along with higher-performance coupled solvers and fluid solvers.
• Motion Generation: Add a research item for Neural Motion Generation.
• Data Pipeline: Combine the agentic skill generation and sim-ready asset/scene-layout generation items into a single in-progress Generative Simulation system (Real2Sim pipeline entry unchanged).
• Models and Training Infrastructure: Replace the VLA framework entry with in-progress APG and motion-control training infrastructure.
• Embodied Tasks: Mark benchmark tasks as in progress; update the third item to call out more deformable-object environment tasks and mark it in progress.

Dependencies: none

Type of change

• Documentation update

Checklist

• I have run the black . command to format the code base.
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works
• Dependencies have been updated, if applicable.

Validation

• black --check passes (627 files unchanged).
• Confirmed roadmap.md is wired into the Sphinx toctree (docs/source/index.rst).
• Full Sphinx docs build was not run: this is a content-only edit (text and table rows) to an existing page with no new directives, code blocks, or cross-references, so build risk is negligible relative to the project's existing warning baseline.

🤖 Generated with Claude Code

[greptile-apps]

Greptile Summary

The PR refreshes the public roadmap to reflect current priorities across simulation, data generation, training infrastructure, and embodied tasks.

• Adds rendering, Newton physics, and neural motion-generation initiatives.
• Consolidates generative-simulation work and replaces the VLA entry with APG and motion-control infrastructure.
• Updates the status and scope of benchmark and deformable-object tasks.

Confidence Score: 4/5

The documentation-only PR appears safe to merge, with one non-blocking clarity issue in the APG roadmap entry.

The roadmap structure and status markers remain valid, but one new user-facing item relies on an unexplained acronym whose meaning is only available elsewhere in the documentation.

Files Needing Attention: docs/source/resources/roadmap.md

Important Files Changed

Filename	Overview
docs/source/resources/roadmap.md	Updates roadmap capabilities and statuses; the new APG entry should expand or link the acronym for roadmap readers.

Prompt To Fix All With AI

### Issue 1
docs/source/resources/roadmap.md:66
**Expand the APG acronym**

The new user-facing roadmap entry uses `APG` without expanding or linking it, requiring readers to search elsewhere to learn that it means Analytic Policy Gradients and understand the planned capability.

```suggestion
| 🚧 | Release Analytic Policy Gradients (APG) training infrastructure. |
```

---

For each issue above, determine whether it is valid and should be fixed. If so, fix it directly.

_{Reviews (1): Last reviewed commit: "Update roadmap items across product area..." | Re-trigger Greptile}

[greptile-apps]

Expand the APG acronym

The new user-facing roadmap entry uses APG without expanding or linking it, requiring readers to search elsewhere to learn that it means Analytic Policy Gradients and understand the planned capability.

Suggested change

	| 🚧 | Release APG training infrastructure. |
	| 🚧 | Release Analytic Policy Gradients (APG) training infrastructure. |

Prompt To Fix With AI

This is a comment left during a code review.
Path: docs/source/resources/roadmap.md
Line: 66

Comment:
**Expand the APG acronym**

The new user-facing roadmap entry uses `APG` without expanding or linking it, requiring readers to search elsewhere to learn that it means Analytic Policy Gradients and understand the planned capability.

```suggestion
| 🚧 | Release Analytic Policy Gradients (APG) training infrastructure. |
```

---

For each issue above, determine whether it is valid and should be fixed. If so, fix it directly.

Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!

[Copilot]

Pull request overview

This PR updates the EmbodiChain Sphinx documentation roadmap (docs/source/resources/roadmap.md) to reflect current priorities across simulation (rendering/physics/motion generation), data pipeline, training infrastructure, and tasks.

Changes:

• Added new in-progress roadmap items for batch rendering optimization and Newton backend + higher-performance solvers.
• Consolidated data-pipeline generative items into a single “generative simulation system” entry and refreshed models/training items to focus on APG + motion-control infrastructure.
• Updated tasks section to mark benchmark and deformable-object task expansion as in progress.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.