C3M is a cross‑session multimodal memory system designed for long‑horizon tasks that must preserve and retrieve evidence across sessions within a limited, query‑blind memory budget. It maintains a bounded active index of source text‑image evidence, using relation‑aware updates to keep safe redundancy while preserving complementary and incompatible records. At query time, budgeted routing selects useful index pages and expands their associated source evidence under a fixed reader budget, creating a compact, provenance‑preserving memory that retains temporal distinctions and source links for reliable downstream reasoning.
arXiv:2607. 29440v1 Announce Type: new Abstract: Long-term multimodal memory must support not only retrieving relevant information but also computing over observations accumulated across interactions.
arXiv:2608.29897v1 Announce Type: new Abstract: Long-horizon agents need a context manager to compress growing interaction histories into a bounded working context, via passive strategies or active s...
GraphMemix introduces a combinatorial‑optimization graph memory framework that organizes long‑term multimodal agent memory as query‑aware evidence forests. It constructs candidate graphs by expanding seed memories through schema and semantic relations, then decouples evidence utility from anchor‑conditioned relation verification to reduce redundancy, and finally optimizes a forest‑format memory context within a maximum evidence budget. Experiments on four benchmarks show significant accuracy gains and a new Pareto frontier between accuracy and lifecycle cost.
GraphMemix introduces a combinatorial‑optimization graph memory framework that constructs query‑aware evidence forests for long‑term multimodal agent memory. It expands seed memories via schema and semantic relations, decouples memory support from relation verification to reduce redundancy, and optimizes a forest‑format context within a maximum evidence budget. Experiments on four benchmarks show significant accuracy gains and a new Pareto frontier between accuracy and lifecycle cost.
FOVEA introduces a cache‑friendly, on‑demand visual evidence adaptation for multimodal speculative decoding, enabling a lightweight draft model to dynamically retrieve a bounded subset of visual memory based on a cumulative‑mass rule. The retrieved visual readout is fused with the draft hidden state via a lightweight gated residual correction, avoiding the insertion of visual tokens into the autoregressive context. Experiments on various vision‑language backbones and benchmarks show that FOVEA improves draft acceptance and speeds up end‑to‑end decoding by up to 2.13× compared to traditional autoregressive decoding.
arXiv:2609.00551v1 Announce Type: cross Abstract: Multimodal memory offers a scalable interface for long-video question answering, but existing methods often retrieve captions, frames, transcripts, s...
arXiv:2605. 28831v2 Announce Type: replace-cross Abstract: Long-horizon memory question answering often requires sparse evidence from heterogeneous histories, including events, object states, visual observations, temporal relations, and causal steps.
MemFuseBench is a benchmark for multi‑source memory fusion that generates source‑tagged observations, evidence‑grounded questions, and adversarial distractors using a Scene‑to‑Sensor pipeline. The proposed MemFuse system preserves source‑level evidence in atomic memory and clusters related events into fused memory via a causal fusion graph, enabling traceable retrieval of dispersed observations. Experiments show that MemFuse outperforms other memory systems across all LLM settings, especially on questions requiring cross‑source evidence fusion.
External memory effectively grounds large language models (LLMs) and vision-language models (VLMs)-based question answering (QA) in relevant multimodal evidence. However, existing memory paradigms represent each memory item in raw text and image forms, so retrieval-based systems must pass the retrieved text or images to the generation LLMs/VLMs, resulting in high token consumption and storage pressure, making it unaffordable for resource-constrained applications.
arXiv:2606. 10572v1 Announce Type: new Abstract: External memory effectively grounds large language models (LLMs) and vision-language models (VLMs)-based question answering (QA) in relevant multimodal evidence.
MemFuse introduces a multi‑source memory fusion system and a corresponding benchmark, MemFuseBench, designed to evaluate agents that must integrate fragmented observations from multiple applications, devices, users, and time points. The benchmark uses a Scene‑to‑Sensor pipeline to generate source‑tagged observations, evidence‑grounded questions, and adversarial distractors, enabling systematic assessment of temporal reasoning, cross‑source evidence fusion, and noise robustness. Experiments show that MemFuse outperforms other memory systems across all evaluated LLM settings, especially on questions requiring cross‑source evidence fusion.
arXiv:2604. 09508v2 Announce Type: replace-cross Abstract: Visual Retrieval-Augmented Generation (VRAG) empowers Vision-Language Models to retrieve and reason over visually rich documents.