RMBench is a simulation benchmark featuring nine robotic manipulation tasks that vary in memory complexity, designed to systematically evaluate how well policies use memory. The authors also introduce Mem-0, a modular policy with explicit memory components that allows controlled ablation studies. Experiments in simulation and the real world reveal memory-related limitations in current policies and show how architectural choices affect memory performance.
arXiv:2609.07047v1 Announce Type: cross Abstract: Robotic manipulation often requires acting on information that is no longer visible, yet Vision-Language-Action policies are usually evaluated when t...
arXiv:2609.22684v1 Announce Type: cross Abstract: Memory-dependent robotic manipulation often requires later actions to use information from earlier interactions. Existing vision-language-action (VLA...
The paper introduces Latent Transition Memory (LTM), a method that stores interaction history as compact latent transitions between consecutive observations for vision‑language‑action models. Implemented as LTM‑π on the π₀.₅ architecture, it uses a frozen latent action encoder, a lightweight adapter, and causal KV caching to enable incremental inference without external reasoning models or memory‑specific annotations. Experiments on RoboMME and LIBERO‑Mem show that LTM‑π achieves state‑of‑the‑art success rates (58.54% and 80.5% respectively), and scaling the dataset or applying a 200‑token memory budget still yields high performance, demonstrating the method’s scalability and efficiency in real‑world robotic tasks.
arXiv:2609.32453v2 Announce Type: replace-cross Abstract: Robotic manipulation is inherently history-dependent, yet most pretrained robotic policies condition on only the current observation or a sho...
arXiv:2609.36595v1 Announce Type: cross Abstract: Visual-memory systems commonly retain or compress past observations. Robot control additionally requires interaction-derived state that no individual...
arXiv:2607. 04434v1 Announce Type: cross Abstract: Generalist robot manipulation policies have advanced rapidly, yet existing benchmarks remain limited in systematically evaluating their capabilities.
The paper introduces 2AM, a system that separates memory and action execution in long‑horizon robot manipulation. 2AM stores task memory exclusively in a multimodal Agent, while a single RGB‑based, stateless Action Model performs motion based on language and optional 2D hints. On the LIBERO‑Mem benchmark, 2AM achieves 76.3% average completion—over 61 points higher than the best baseline—demonstrating that agent‑side memory and precise steering of the Action Model can substantially improve performance.
The paper introduces 2AM, a system that keeps task memory solely within a multimodal Agent while using a single RGB‑based, stateless Action Model to execute motions. By compiling interaction history into subtask language and optional 2D grasp/place/move hints, the Agent steers the Action Model, which is trained to tolerate imperfect guidance through dropout, noise, and jitter. On the LIBERO‑Mem benchmark, 2AM achieves 76.3% average completion without depth, geometry, or planners, vastly outperforming the best baseline.
arXiv:2604. 18933v2 Announce Type: replace-cross Abstract: Robotic manipulation tasks exhibit varying memory requirements, ranging from Markovian tasks that require no memory to non-Markovian tasks that demand in-context memorization of historical information within a single trial or in-context adaptation based on the outcomes of multiple past trials.
arXiv:2508. 21378v2 Announce Type: replace-cross Abstract: Large language models (LLMs) demonstrate remarkable capabilities in reasoning and code generation, enabling robotic manipulation to be initiated with just a single instruction.
arXiv:2603. 24576v2 Announce Type: replace-cross Abstract: Robots often observe information that determines a future action long before that action is executed.
PREPING introduces a proposer-guided framework for constructing agent memory before encountering target tasks, using self-generated synthetic practice. The method employs a proposer memory to shape synthetic tasks, a solver to execute them, and a validator to filter useful trajectories, thereby avoiding redundancy and infeasibility. Experiments on AppWorld, BFCL v3, and MCP-Universe demonstrate that PREPING outperforms a no-memory baseline and rivals playbook-based approaches while reducing deployment costs by 2.99× and 2.23× respectively.