CVAIJun 10

Task-Aware Structured Memory for Dynamic Multi-modal In-Context Learning

arXiv:2606.11853v19.8h-index: 4
Predicted impact top 50% in CV · last 90 daysOriginality Incremental advance
AI Analysis

This work addresses the scalability bottleneck of MLLMs due to finite context windows and growing KV cache costs, offering a practical solution for efficient multi-modal ICL without retraining.

TASM introduces a training-free framework for dynamic multi-modal in-context learning that compresses long sequences into task-aware, structure-preserving memories, achieving high performance under heavy compression while balancing efficiency and adaptability.

Multi-modal large language models (MLLMs) depend on in-context learning (ICL) for rapid task adaptation, but their scalability is severely limited by finite context windows and the growing cost of key-value (KV) caches in long multi-modal sequences. Existing memory compression approaches typically rely on rigid token removal or sample-dependent importance estimation, which introduces bias, disrupts semantic structure, particularly for visual representations, and yields static memories that cannot adapt to new queries. We introduce TASM (Task-Aware Structured Memory), a training-free framework that addresses these limitations through task-aware, structure-preserving, and dynamically accessible memory construction. TASM employs task-vector guided compression to replace sample-specific signals with a task-level direction that captures shared relevance across demonstrations. To preserve the underlying manifold, it applies semantics-aware token merging via bipartite graph matching, aggregating tokens without destructive pruning. Finally, TASM structures memory into a hierarchy comprising a compact Core Memory and a Latent Bank, facilitating query-adaptive dynamic retrieval. Evaluations confirm TASM maintains high performance under heavy compression, effectively balancing efficiency with adaptability.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes