Retrieved but Not Delivered: Multimodal Memory Delivery for Long-Term Agents

arXiv:2609.32590 · cs.CV, cs.AI, cs.CL, cs.IR · Submitted 2026-09-26 · Read on arXiv

cs.CV, cs.AI, cs.CL, cs.IR

Submitted: 2026-09-26

Updated: 2026-09-26

Project page: https://avalon-s.github.io/DeliverMem

Terminology

• Benchmark — 83× in this paper · explained in 2 episodes
• Accuracy — 22× in this paper · explained in 1 episode
• Trace — 15× in this paper · explained in 2 episodes
• Addressability — 10× in this paper · explained in 1 episode
• Agent Memory — 9× in this paper · explained in 1 episode
• Temporal Reasoning — 9× in this paper · explained in 2 episodes
• Abstention — 5× in this paper · explained in 1 episode
• Embeddings — 3× in this paper · explained in 2 episodes
• Generative AI — 3× in this paper · explained in 2 episodes
• Ablation — 2× in this paper · explained in 1 episode
• Benchmarking — 2× in this paper · explained in 3 episodes
• AI Agents — 1× in this paper · explained in 2 episodes
• Adam — 1× in this paper · explained in 2 episodes
• Agent Interactions — 1× in this paper · explained in 1 episode
• Decode — 1× in this paper · explained in 2 episodes
• Grounding — 1× in this paper · explained in 3 episodes
• Mixture of Experts — 1× in this paper · explained in 2 episodes
• Reinforcement Learning — 1× in this paper · explained in 4 episodes
• Semantic Understanding — 1× in this paper · explained in 2 episodes
• Transformer — 1× in this paper · explained in 2 episodes
• Vision-language models — 1× in this paper · explained in 2 episodes
• Visual Grounding — 1× in this paper · explained in 2 episodes

Sources

• Argus-Retriever: Vision-LLM Late-Interaction Retrieval with Region-Aware Query-Conditioned MoE for Visual Document Retrieval
• Qwen3-VL Technical Report
• Qwen2.5-VL Technical Report
• SMMBench: A Benchmark for Source-Distributed Multimodal Agent Memory
• Naive Visual Memory is Not Enough: A Failure-Mode Study of GUI Agents
• MemArbiter: Decision-Time Memory Arbitration for Long-Horizon LLM Agents
• Memory-T1: Reinforcement Learning for Temporal Reasoning in Multi-session Agents
• M2A: Multimodal Memory Agent with Dual-Layer Hybrid Memory for Long-Term Personalized Interactions
• Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
• MemEye: A Visual-Centric Evaluation Framework for Multimodal Agent Memory
• M$^3$Exam: Benchmarking Multimodal Memory for Realistic User-Agent Interactions
• Enabling Chatbots with Eyes and Ears: An Immersive Multimodal Conversation System for Dynamic Interactions
• Presentation, Not Mechanism: A Render Confound in Deprecation-Aware Memory Evaluation
• V-Mem: Modality-Routed Retrieval for Long-Term Multimodal Agentic Memory
• MemOS: An Operating System for Memory-Augmented Generation (MAG) in Large Language Models
• MemVerse: Multimodal Memory for Lifelong Learning Agents
• Seeing, Listening, Remembering, and Reasoning: A Multimodal Agent with Long-Term Memory
• MMA: Multimodal Memory Agent
• Agent Memory: Characterization and System Implications of Stateful Long-Horizon Workloads
• MemGPT: Towards LLMs as Operating Systems

Related papers

• Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
• AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
• Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
• MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
• TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
• A Survey on Efficient Vision-Language-Action Models