10 papers · 1 filter
When Plans Change, Remember: Replanning-Guided Visual Memory for Vision-Language-Action Models
Victoria Smelova, Daniil Kazachkov, Georgiy Kazantsev +6
Long-horizon, multi-stage Vision-Language-Action (VLA) policies often require visual information that is no longer present in the current observation. Retaining the complete observ…
MIKASA-Robo-VLA: Benchmarking Memory in VLA Models for Long-Horizon Manipulation
Egor Cherepanov, Nikita Kachaev, Aleksandr I. Panov +1
Vision-language-action policies often see only one or a few recent frames, which makes it difficult to evaluate how they use information that disappears during a task. We introduce…
ALER: Adaptive Learnable Experience Rewriting for Reinforcement Learning
Oleg Shchendrigin, Egor Cherepanov, Aleksandr I. Panov +1
In partially observable reinforcement learning (RL), a later observation can make stored information obsolete or change what it implies for the next decision. Memory architectures…
Leveraging Offline Supervision for Efficient and Generalizable Reinforcement Learning in Large-Scale Vision-Language-Action Models
Dmitriy Poyarkov, Aleksei Staroverov, Aleksandr I. Panov
It is commonly observed that online reinforcement learning (RL) produces better-performing strategies than offline methods across a broad range of performance measures. In particul…
VLA Grounder: Language-Conditioning Space Optimization for Black-Box VLA Models
Damir Shodiev, Aleksei Staroverov, Nikita Kachaev +2
Vision-Language-Action (VLA) models are commonly treated as end-to-end action policies conditioned on natural-language task descriptions. In practice, however, their behavior often…
Imagine to Ensure Safety in Hierarchical Reinforcement Learning
Gregory Gorbov, Artem Latyshev, Aleksandr I. Panov
This work investigates the safe exploration problem in reinforcement learning, where an agent must maximize cumulative performance while simultaneously satisfying safety constraint…
Does VLA Even Know the Basics? Measuring Commonsense and World Knowledge Retention in Vision-Language-Action Models
Nikita Kachaev, Andrey Moskalenko, Matvey Skripkin +10
Embodied Vision-Language-Action (VLA) models are typically obtained by fine-tuning powerful pretrained VLMs on robotics data, yet it is unclear how much commonsense and factual kno…
VLA: On Recurrent Memory for Partially Observable Manipulation in VLA Models
Egor Cherepanov, Nikita Kachaev, Daniil Zelezetsky +6
Vision-language-action (VLA) models predict chunks of future actions from the current observation, an assumption that fails under partial observability, where decisions depend on i…
Memory Retention Is Not Enough to Master Memory Tasks in Reinforcement Learning
Oleg Shchendrigin, Egor Cherepanov, Alexey K. Kovalev +1
Effective decision-making in the real world depends on memory that is both stable and adaptive: environments change over time, and agents must retain relevant information over long…
KAGE-Bench: Fast Known-Axis Visual Generalization Evaluation for Reinforcement Learning
Egor Cherepanov, Daniil Zelezetsky, Alexey K. Kovalev +1
Pixel-based reinforcement learning agents often fail under purely visual distribution shift even when latent dynamics and rewards are unchanged, but existing benchmarks entangle mu…