# PhysMamba: Selective State Space Models as Learned Articulated Body Simulators

arXiv: https://arxiv.org/abs/2610.04014 · HTML: https://mlsanity.com/paper/2610.04014/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Video, motion and generative Mamba models, one of 10 directions in State-space sequence models (S4, Mamba).
- Builds on 9 of its 16 references that are in the map.
- First posted on arXiv 2 Oct 2026.

## Placed in

- State-space sequence models (S4, Mamba) / [Video, motion and generative Mamba models](https://mlsanity.com/ssm/video-motion-and-generative-mamba-models/)

## Builds on (most cited first, 9 of 9)

- [Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality](https://mlsanity.com/paper/2405.21060/) (2024)
- [AMASS: Archive of Motion Capture as Surface Shapes](https://mlsanity.com/paper/1904.03278/) (2019)
- [Graph networks as learnable physics engines for inference and control](https://mlsanity.com/paper/1806.01242/) (2018)
- [Perpetual Humanoid Control for Real-time Simulated Avatars](https://mlsanity.com/paper/2305.06456/) (2023)
- [Universal Humanoid Motion Representations for Physics-Based Control](https://mlsanity.com/paper/2310.04582/) (2023)
- [SimPoE: Simulated Character Control for 3D Human Pose Estimation](https://mlsanity.com/paper/2104.00683/) (2021)
- [Differentiable Dynamics for Articulated 3d Human Motion Reconstruction](https://mlsanity.com/paper/2205.12256/) (2022)
- [Learned Neural Physics Simulation for Articulated 3D Human Pose Reconstruction](https://mlsanity.com/paper/2410.12023/) (2024)
- [A Generalizable Physics-Enhanced State Space Model for Long-Term Dynamics Forecasting in Complex Environments](https://mlsanity.com/paper/2507.10792/) (2025)

## Abstract (arXiv)

We introduce the first learned articulated body simulator based on a selective state space model (SSM), called PhysMamba. PhysMamba predicts next-frame full-body state from position, rotation, and joint-action history, without velocity inputs. We compare four architectures under partial- and full-observation inputs and three training protocols. The from-scratch rollout training protocol gives Mamba2 strong short- and mid-horizon accuracy under partial observation (s10 = 43 mm, 2/50 diverged), while the two-stage teacher-based rollout protocol stabilizes GRU but fails for Mamba2. With CUDA graph compilation, Mamba2 reaches 0.107 ms per frame (9,334 FPS) on an H100 GPU, within 1.1× of GRU's un-compiled throughput, adding under 1% latency to a 30 Hz HMR pipeline and enabling integration as a differentiable physics module for video-based mesh recovery.