STaR-KV: Spatio-Temporal Adaptive Re-weighting for KV Cache Compression in GUI Vision-Language Models
Yuhang Han , Wenzheng Yang , Yujie Chen , Xiangqi Jin , Yaojie Zhang , Siteng Huang , Linfeng Zhang
- 🏛 Institutions
- SJTU , HKUST(GZ) , ZJU
- 📅 Date
- June 1, 2026
- 📑 Publisher
- arXiv
- 💻 Env
- General GUI
- 🔑 Keywords
TLDR
STaR-KV is a training-free KV cache compression framework for GUI vision-language models, whose cache grows linearly with interaction steps. It combines subspace-aware spatial scoring with temporal-stability re-weighting to retain salient tokens, reducing memory while preserving accuracy across GUI agent benchmarks.
Related papers (24)
- GUI-KV: Efficient GUI Agents via KV Cache with Spatio-Temporal AwarenessOctober 1, 2025 · arXiv
- PhoneCLI: From App Interfaces to Callable Commands for Mobile AgentsSeptember 28, 2026 · arXiv
- GUIFlow: Temporal Redundancy-Aware Incremental Inference for GUI AgentsOctober 26, 2026 · arXiv
- When to Compile a Computer-Use Agent? Measuring Payback and Making Compilation Decisions for Token EfficiencyOctober 2, 2026 · arXiv
- cua-speedrun: Standardized Benchmarking of the Speed of Computer-Use AgentsSeptember 30, 2026 · arXiv
- Relevance Does Not Imply Applicability: Experience Activation for Personal GUI AgentsSeptember 27, 2026 · arXiv
- TRACE: Trajectory-robust Admission with Evidence Ordering for Efficient GUI AgentsSeptember 9, 2026 · arXiv
- Efficient GUI Agents: A Systems Survey of Observation, Memory, Action, and Runtime OptimizationSeptember 2, 2026 · arXiv
- Step-level Optimization for Efficient Computer-use AgentsApril 29, 2026 · arXiv
- UI-Zoomer: Uncertainty-Driven Adaptive Zoom-In for GUI GroundingApril 15, 2026 · arXiv
- Zoom to Essence: Trainless GUI Grounding by Inferring upon Interface ElementsMarch 15, 2026 · arXiv
- Trifuse: Enhancing Attention-Based GUI Grounding via Multimodal FusionFebruary 6, 2026 · arXiv
- Darwinian Memory: A Training-Free Self-Regulating Memory System for GUI Agent EvolutionJanuary 30, 2026 · arXiv
- MVP: Multiple View Prediction Improves GUI GroundingDecember 9, 2025 · arXiv
- Zoom in, Click out: Unlocking and Evaluating the Potential of Zooming for GUI GroundingDecember 5, 2025 · arXiv
- Tandem: Serving Hybrid Linear-Attention GUI Agents on Mobile NPUsOctober 19, 2026 · arXiv
- STP-GUI: Spatially-Aware Token Pruning for Efficient Web-GUI AgentsOctober 14, 2026 · arXiv
- MHVF: Training-Free GUI Grounding via Verification-Guided Multi-Hypothesis SearchOctober 13, 2026 · arXiv
- CUAWright: A Minimal Unified Interface for Digital AgentsOctober 2, 2026 · arXiv
- In-Context Teacher–Student Guidance for Open-Weight Browser AgentsSeptember 2026 · KDMiLe 2026
- Action Conditioned Bisimulation For GUI Agent MemorySeptember 30, 2026 · arXiv
- Neuro-Symbolic Computer Use: Learning Reusable Policies for Reliable and Efficient ExecutionSeptember 29, 2026 · arXiv
- Absorbed in Inertia: Activation Analysis for Computer-Use AgentsSeptember 29, 2026 · arXiv
- PastForward: Faster On-Device GUI Agents via Computational Experience ReuseSeptember 26, 2026 · arXiv