ReMem-VLA: Empowering Vision-Language-Action Model with Memory via Dual-Level Recurrent Queries
This addresses memory limitations in robot control models, though it appears incremental as it builds on existing VLA frameworks with memory enhancements.
The paper tackles the problem of vision-language-action models lacking historical context for robot control by introducing ReMem-VLA, which uses dual-level recurrent queries for memory; it significantly outperforms memory-free baselines and MemoryVLA on memory-dependent tasks.
Vision-language-action (VLA) models for closed-loop robot control are typically cast under the Markov assumption, making them prone to errors on tasks requiring historical context. To incorporate memory, existing VLAs either retrieve from a memory bank, which can be misled by distractors, or extend the frame window, whose fixed horizon still limits long-term retention. In this paper, we introduce ReMem-VLA, a Recurrent Memory VLA model equipped with two sets of learnable queries: frame-level recurrent memory queries for propagating information across consecutive frames to support short-term memory, and chunk-level recurrent memory queries for carrying context across temporal chunks for long-term memory. These queries are trained end-to-end to aggregate and maintain relevant context over time, implicitly guiding the model's decisions without additional training or inference cost. Furthermore, to enhance visual memory, we introduce Past Observation Prediction as an auxiliary training objective. Through extensive memory-centric simulation and real-world robot experiments, we demonstrate that ReMem-VLA exhibits strong memory capabilities across multiple dimensions, including spatial, sequential, episodic, temporal, and visual memory. ReMem-VLA significantly outperforms memory-free VLA baselines $Ï$0.5 and OpenVLA-OFT and surpasses MemoryVLA on memory-dependent tasks by a large margin.