mlsanityData to 14 Oct 2026

FlowMM: Cross-Modal Information Flow Guided KV Cache Merging for Efficient Multimodal Context Inference

arXiv 2511.05534 ↗First posted 29 Oct 2025cs.CLcited by 6

On the map: Efficient LLM inference and serving → KV cache optimization

Abstract (arXiv)

Traditional KV cache eviction strategies, which discard less critical KV-pairs based on attention scores, often degrade generation quality, causing context loss or hallucinations. Recent efforts shift toward KV merging, merging eviction tokens with retention tokens based on similarity. However, in multimodal scenarios, distributional biases across modality tokens and attentional biases in cross-modal interactions limit its effectiveness. This work introduces FlowMM, an adaptive framework for cross-modal information flow-guided multimodal KV cache merging. FlowMM leverages cross-modal information flow to dynamically apply layer-specific merging strategies, capturing modality-specific patterns while preserving contextual integrity. Furthermore, we introduce a sensitivity-adaptive token matching mechanism that jointly evaluates token similarity and task-critical sensitivity, merging low-risk tokens while safeguarding high-sensitivity ones. Extensive experiments across diverse leading MLLMs show that FlowMM reduces KV cache memory by 80% to 95% and decoding latency by 1.3-1.8x, while maintaining competitive task performance.

Key facts. Cited by 6 papers in the map (data through 14 Oct 2026). Placed in KV cache optimization, one of 12 directions in Efficient LLM inference and serving. Builds on 37 of its 43 references that are in the map. First posted on arXiv 29 Oct 2025.

Builds on

37 of 43 references.

Closest ancestors · 7

Same direction, newest first

  1. 2024Model Tells You Where to Merge: Adaptive KV Cache Merging for LLMs on Long-Context Tasks11 Jul
  2. 2024LOOK-M: Look-Once Optimization in KV Cache for Efficient Multimodal Long-Context Inference26 Jun
  3. 2024MiniCache: KV Cache Compression in Depth Dimension for Large Language Models23 May
  4. 2024SnapKV: LLM Knows What You are Looking for Before Generation22 Apr
  5. 2024GEAR: An Efficient KV Cache Compression Recipe for Near-Lossless Generative Inference of LLM8 Mar
  6. 2024No Token Left Behind: Reliable KV Cache Compression via Importance-Aware Mixed Precision Quantization28 Feb
  7. 2023H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models24 Jun

Other references · 30

  1. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  2. 2023Visual Instruction Tuningcited by 7,869
  3. 2023BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Modelscited by 6,137
  4. 2025Qwen2.5-VL Technical Reportcited by 5,587
  5. 2022Flamingo: a Visual Language Model for Few-Shot Learningcited by 4,386
  6. 2024Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolutioncited by 4,234
  7. 2023Efficient Streaming Language Models with Attention Sinkscited by 2,186
  8. 2022GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformerscited by 1,919
  9. 2024Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scalingcited by 1,599
  10. 2022SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Modelscited by 1,386
  11. 2023AWQ: Activation-aware Weight Quantization for LLM Compression and Accelerationcited by 1,298
  12. 2023SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shotcited by 1,121
18 more
  1. 2023A Simple and Effective Pruning Approach for Large Language Modelscited by 800
  2. 2019ALFRED: A Benchmark for Interpreting Grounded Instructions for Everyday Taskscited by 783
  3. 2024An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Modelscited by 642
  4. 2024KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cachecited by 637
  5. 2022ZeroQuant: Efficient and Affordable Post-Training Quantization for Large-Scale Transformerscited by 526
  6. 2024LLaVA-PruMerge: Adaptive Token Reduction for Efficient Large Multimodal Modelscited by 402
  7. 2024MobileVLM V2: Faster and Stronger Baseline for Vision Language Modelcited by 174
  8. 2024Fit and Prune: Fast and Training-free Visual Token Pruning for Multi-modal Large Language Modelscited by 126
  9. 2023MobileVLM : A Fast, Strong and Open Vision Language Assistant for Mobile Devicescited by 122
  10. 2018Learning to Describe Differences Between Pairs of Similar Imagescited by 114
  11. 2024MileBench: Benchmarking MLLMs in Long Contextcited by 88
  12. 2024Cross-modal Information Flow in Multimodal Large Language Modelscited by 67
  13. 2024MADTP: Multimodal Alignment-Guided Dynamic Token Pruning for Accelerating Vision-Language Transformercited by 54
  14. 2019Expressing Visual Relationships via Languagecited by 49
  15. 2025Disentangled Knowledge Tracing for Alleviating Cognitive Biascited by 12
  16. 2025Cuff-KT: Tackling Learners' Real-time Learning Pattern Adjustment via Tuning-Free Knowledge State Guided Model Updatingcited by 5
  17. 2024FedCFA: Alleviating Simpson's Paradox in Model Aggregation with Counterfactual Federated Learningcited by 4
  18. 2025DKT2: Revisiting Applicable and Comprehensive Knowledge Tracing in Large-Scale Datacited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 6.

6 papers.

  1. 2026RetentiveKV: State-Space Memory for Uncertainty-Aware Multimodal KV Cache Evictioncited by 3
  2. 2026ELSA3D: Elastic Semantic Anchoring for Unified 3D Understanding and Generationcited by 2
  3. 2026Beyond Student: An Asymmetric Network for Neural Network Inheritancecited by 1
  4. 2026KVSlimmer: Theoretical Insights and Practical Optimizations for Asymmetric KV Mergingcited by 0
  5. 2026Reducing Peak Memory Usage for Modern Multimodal Large Language Model Pipelinescited by 0
  6. 2026A3B2: Adaptive Asymmetric Adapter for Alleviating Branch Bias in Vision-Language Image Classification with Few-Shot Learningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025MEDA: Dynamic KV Cache Allocation for Efficient Multimodal Long-Context Inference
  2. 2026Hierarchical Adaptive Eviction for KV Cache Management in Multimodal Language Models
  3. 2026UniCache: Task- and Type-Aware KV Cache Compression for Unified Multimodal Models
  4. 2025MadaKV: Adaptive Modality-Perception KV Cache Eviction for Efficient Multimodal Long-Context Inference
  5. 2026Last But Not Least: Boundary Attention CalibratiON for Multimodal KV Cache Compression
  6. 2025SmallKV: Small Model Assisted Compensation of KV Cache Compression for Efficient LLM Inference
  7. 2025Mixing Importance with Diversity: Joint Optimization for KV Cache Compression in Large Vision-Language Models
  8. 2026Omni-Flow: A Unified Workflow Orchestration and Distributed KV Cache Sharing Framework for Multimodal Inference