Cover Image

OxyGen: Unified KV Cache Management for VLA Inference under Multi-Task Parallelism

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Li, Xiangyu, Tang, Huaizhi, Ding, Xin, Wang, Weijun, Cao, Ting, Liu, Yunxin
Format:	Preprint
Published:	2026
Subjects:	Robotics Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Sci-VLA: Agentic VLA Inference Plugin for Long-Horizon Tasks in Scientific Experiments
by: Pang, Yiwen, et al.
Published: (2026)

Vec-LUT: Vector Table Lookup for Parallel Ultra-Low-Bit LLM Inference on Edge Devices
by: Li, Xiangyu, et al.
Published: (2025)

KEEP: A KV-Cache-Centric Memory Management System for Efficient Embodied Planning
by: Yang, Zebin, et al.
Published: (2026)

KVShare: An LLM Service System with Efficient and Effective Multi-Tenant KV Cache Reuse
by: Yang, Huan, et al.
Published: (2025)

FUTURE-VLA: Forecasting Unified Trajectories Under Real-time Execution
by: Fan, Jingjing, et al.
Published: (2026)

Qwen-VLA: Unifying Vision-Language-Action Modeling across Tasks, Environments, and Robot Embodiments
by: Wang, Qiuyue, et al.
Published: (2026)

LoHoVLA: A Unified Vision-Language-Action Model for Long-Horizon Embodied Tasks
by: Yang, Yi, et al.
Published: (2025)

BioProVLA-Agent: An Affordable, Protocol-Driven, Vision-Enhanced VLA-Enabled Embodied Multi-Agent System with Closed-Loop-Capable Reasoning for Biological Laboratory Manipulation
by: Du, Zhaohui, et al.
Published: (2026)

VacuumVLA: Boosting VLA Capabilities via a Unified Suction and Gripping Tool for Complex Robotic Manipulation
by: Zhou, Hui, et al.
Published: (2025)

Towards Deploying VLA without Fine-Tuning: Plug-and-Play Inference-Time VLA Policy Steering via Embodied Evolutionary Diffusion
by: Li, Zhuo, et al.
Published: (2025)

HiMoE-VLA: Hierarchical Mixture-of-Experts for Generalist Vision-Language-Action Policies
by: Du, Zhiying, et al.
Published: (2025)

Beyond Task Success: Behavioral and Representational Diagnostics for WAM and VLA
by: Mai, Hung, et al.
Published: (2026)

Parallels Between VLA Model Post-Training and Human Motor Learning: Progress, Challenges, and Trends
by: Xiang, Tian-Yu, et al.
Published: (2025)

VLA-RAIL: A Real-Time Asynchronous Inference Linker for VLA Models and Robots
by: Zhao, Yongsheng, et al.
Published: (2025)

MetaVLA: Unified Meta Co-training For Efficient Embodied Adaption
by: Li, Chen, et al.
Published: (2025)

Efficient Remote KV Cache Reuse with GPU-native Video Codec
by: Mi, Liang, et al.
Published: (2026)

Task2Morph: Differentiable Task-inspired Framework for Contact-Aware Robot Design
by: Cai, Yishuai, et al.
Published: (2024)

Enhancing Lifelong Multi-Agent Path Finding with Cache Mechanism
by: Tang, Yimin, et al.
Published: (2025)

RAILGUN: A Unified Convolutional Policy for Multi-Agent Path Finding Across Different Environments and Tasks
by: Tang, Yimin, et al.
Published: (2025)

DroneVLA: VLA based Aerial Manipulation
by: Mehboob, Fawad, et al.
Published: (2026)

SafeGen-LLM: Enhancing Safety Generalization in Task Planning for Robotic Systems
by: Fan, Jialiang, et al.
Published: (2026)

WholeBodyVLA: Towards Unified Latent VLA for Whole-Body Loco-Manipulation Control
by: Jiang, Haoran, et al.
Published: (2025)

RoboPARA: Dual-Arm Robot Planning with Parallel Allocation and Recomposition Across Tasks
by: Duan, Shiying, et al.
Published: (2025)

Caching-Augmented Lifelong Multi-Agent Path Finding
by: Tang, Yimin, et al.
Published: (2024)

UniVLA: Learning to Act Anywhere with Task-centric Latent Actions
by: Bu, Qingwen, et al.
Published: (2025)

Multi-Task Interactive Robot Fleet Learning with Visual World Models
by: Liu, Huihan, et al.
Published: (2024)

Decomposition-based Hierarchical Task Allocation and Planning for Multi-Robots under Hierarchical Temporal Logic Specifications
by: Luo, Xusheng, et al.
Published: (2023)

RaceVLA: VLA-based Racing Drone Navigation with Human-like Behaviour
by: Serpiva, Valerii, et al.
Published: (2025)

HyperVLA: Efficient Inference in Vision-Language-Action Models via Hypernetworks
by: Xiong, Zheng, et al.
Published: (2025)

ContextVLA: Vision-Language-Action Model with Amortized Multi-Frame Context
by: Jang, Huiwon, et al.
Published: (2025)

WorldVLA: Towards Autoregressive Action World Model
by: Cen, Jun, et al.
Published: (2025)

RoboGene: Boosting VLA Pre-training via Diversity-Driven Agentic Framework for Real-World Task Generation
by: Zhang, Yixue, et al.
Published: (2026)

Reinforcement Learning Enabled Adaptive Multi-Task Control for Bipedal Soccer Robots
by: Zhang, Yulai, et al.
Published: (2026)

SpatialVLA: Exploring Spatial Representations for Visual-Language-Action Model
by: Qu, Delin, et al.
Published: (2025)

VLA-RL: Towards Masterful and General Robotic Manipulation with Scalable Reinforcement Learning
by: Lu, Guanxing, et al.
Published: (2025)

Sample-Efficient Robot Skill Learning for Construction Tasks: Benchmarking Hierarchical Reinforcement Learning and Vision-Language-Action VLA Model
by: Hu, Zhaofeng, et al.
Published: (2025)

Eva-VLA: Evaluating Vision-Language-Action Models' Robustness Under Real-World Physical Variations
by: Liu, Hanqing, et al.
Published: (2025)

DeeR-VLA: Dynamic Inference of Multimodal Large Language Models for Efficient Robot Execution
by: Yue, Yang, et al.
Published: (2024)

Shifting Uncertainty to Critical Moments: Towards Reliable Uncertainty Quantification for VLA Model
by: Tang, Yanchuan, et al.
Published: (2026)

MoMaGen: Generating Demonstrations under Soft and Hard Constraints for Multi-Step Bimanual Mobile Manipulation
by: Li, Chengshu, et al.
Published: (2025)