ReCAT: Remember, Count, and Time: Structured Recurrent Memory for Robot Manipulation
cs.RO, cs.AI, cs.CV, cs.LG
Submitted: 2026-09-28
Updated: 2026-09-28
Project page: https://intuitiverobots.github.io/ReCAT
Terminology
Sources
- ReMem-VLA: Empowering Vision-Language-Action Model with Memory via Dual-Level Recurrent Queries
- Gated Delta Networks: Improving Mamba2 with Delta Rule
- DAM-VLA: Decoupled Asynchronous Multimodal Vision Language Action model
- RoboSSM: Scalable In-context Imitation Learning via State-Space Models
- DSSP: Diffusion State Space Policy with Full-History Encoding
- Chronos: A Physics-Informed Full-History Framework for Non-Markovian Long-Horizon Manipulation
- $\mu$VLA: On Recurrent Memory for Partially Observable Manipulation in VLA Models
- BPP: Long-Context Robot Imitation Learning by Focusing on Key History Frames
- SimpleMemVLA: A Simple but Effective Native-Video Memory for Vision-Language-Action Models
- MEM: Multi-Scale Embodied Memory for Vision Language Action Models
- Memory Retrieval in Visuomotor Policies for Long-Horizon Robot Control
- Gated Memory Policy: In-Context Memorization and Adaptation
- DiSPo: Diffusion-SSM based Policy Learning for Coarse-to-Fine Action Discretization
- AnoleVLA: Lightweight Vision-Language-Action Model with Deep State Space Models for Mobile Manipulation
- RoboTTT: Context Scaling for Robot Policies
- Mamba: Linear-Time Sequence Modeling with Selective State Spaces
- Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality
- Zoology: Measuring and Improving Recall in Efficient Language Models
- Kimi Linear: An Expressive, Efficient Attention Architecture
- DynaFLIP: Rethinking Robotics Perception via Tri-Modal-Dynamics Guided Representation
Related papers
- FMT x: An Efficient and Asymptotically Optimal Extension of the Fast Marching Tree for Dynamic Replanning
- MPCFormer: A physics-informed data-driven approach for explainable socially-aware autonomous driving
- RoboLab: A High-Fidelity Simulation Benchmark for Analysis of Task Generalist Policies
- HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments
- APT: Action Expert Pretraining Improves Instruction Generalization of Vision-Language-Action Policies
- Fine-tuning is Not Enough: A Parallel Framework for Collaborative Imitation and Reinforcement Learning in End-to-end Autonomous Driving