5arXiv cs.CL (Computation and Language)·
Supra Cognitive Modes: routed architecture for agent memory across factual, relational, and synthesis workloads
A new arXiv preprint introduces Supra Cognitive Modes (SCM), an agent memory architecture that routes queries to specialized retrieval and synthesis pipelines over a shared ingest substrate combining dense embeddings, knowledge graph triples, and fact-version metadata. A frozen semantic classifier dispatches queries among lexical/dense lookup, multi-hop graph reasoning, and long-form synthesis modes. The system is evaluated on three agent memory benchmarks—LoCoMo (84.87% factoid), MemoryAgentBench (61.49%), and LongMemEval (86.00%)—though the authors note that causal routing effects, efficiency gains, and statistical significance remain unestablished.
Related guides (3)
Related events (8)
Systematic evaluation of 12 agent memory systems from a data management perspective
A new arXiv preprint proposes an analytical framework decomposing agent memory into four core modules—representation/storage, extraction, retrieval/routing, and maintenance—and evaluates 12 representative memory systems across five benchmark workloads spanning 11 datasets. The study finds no single architecture dominates across scenarios; effectiveness depends on alignment between memory structure and workload bottleneck. Fine-grained ablation studies quantify effects on retrieval precision, update correctness, and long-horizon stability, and reveal that localized maintenance is more cost-efficient than global reorganization. Code is publicly released.
Cognitive-structured Multimodal Agent with Episodic Visual Memory for long-horizon dialogue
Researchers propose a Cognitive-structured Multimodal Agent (CMA) that externalizes visual context into an Episodic Visual Memory to avoid token explosion in long multimodal dialogues. The architecture includes a Perceptual Abstraction Engine, Cognitive Retrieval Engine, and Multimodal Executive Controller, trained with reinforcement learning on a synthetically generated multi-turn dataset. An 8B parameter agent achieves 91.4% retrieval accuracy over 20-turn sessions, outperforming 32B baselines by 8.2% while nearly halving inference time. The authors also release CMA-Harness, a tool-augmented deployment integrating persistent memory, web access, and image generation tools with OpenAI-compatible serving.
Infini Memory: Topic-structured persistent memory architecture for long-term LLM agents
Researchers propose Infini Memory, a persistent memory architecture for LLM agents that organizes memory as topic-structured documents rather than isolated records or summaries. New observations are staged in a buffer and periodically consolidated, while retrieval uses iterative agentic tool calls instead of a single lookup step. The system achieves 64.7% on MemoryAgentBench, with ablations showing complementary gains from topic-structured maintenance and iterative evidence inspection.
GitOfThoughts: Git-based agent memory substrate with sobering findings on memory utility for novel problems
Researchers introduce GitOfThoughts, a system that stores LLM reasoning trees as git repositories, enabling replayable, auditable, and mergeable agent memory at low engineering cost. Across five memory substrates (none, markdown, vector, graph, git), two benchmarks, and two model scales with pre-registered replications, the paper finds that no memory format reliably improves accuracy on novel problems. Memory only helps above a 'copyability threshold' (similarity >~0.8), where retrieved cases are near-duplicates of the current problem — and even then, the gain is answer retrieval rather than method transfer. The paper also documents a retracted result and refuted hypothesis, modeling a rigorous evaluation standard.
M³Exam: Benchmark for Multimodal Memory in Realistic User-Agent Interactions
Researchers introduce M³Exam, a query-centric multimodal conversational memory benchmark designed to evaluate language agents on realistic user-agent interactions, including cross-modal grounding and implicit information inference. Existing benchmarks are critiqued for assuming sparse visuals and human-human interaction formats. The paper also proposes M³Proctor, a companion memory method that detects query modality bias and retrieves raw visual sources on demand, achieving 13% accuracy improvement while reducing index-construction time and retrieved tokens by over 70%.
InMind benchmark exposes implicit-association blind spot in agent long-term memory systems
Researchers introduce InMind, a 125-task expert-verified benchmark targeting a failure mode in agent long-term memory: facts stored in memory are not retrieved when the query lacks surface-level textual overlap with the relevant memory (e.g., a nut allergy should affect a macaron recommendation but shares no keywords). Testing six vector, graph, and agentic memory systems reveals a stark gap — backbone models answer 84% of indirect queries when the relevant memory is placed in context, but retrieval systems surface the correct answer at most 14.4% of the time despite near-perfect on-demand recall. The study isolates the failure to the query-conditioned retrieval interface itself, framing routing — deciding which facts must remain visible — as the core open problem.
MemDreamer: Hierarchical graph memory and agentic retrieval for long video understanding
MemDreamer is a plug-and-play framework that decouples perception and reasoning for long-video understanding by incrementally building a three-tier Hierarchical Graph Memory capturing spatiotemporal and causal relations. During inference, a reasoning model uses an Observation-Reason-Action loop with agentic tool-augmented retrieval to navigate the memory graph, constraining the context window to 2% of full-context ingestion while achieving a 12.5-point absolute accuracy gain. The system reaches SOTA on four benchmarks, narrowing the gap with human experts to 3.7 points. The authors also report a strong linear correlation between logical reasoning performance and long-video understanding, proposing agentic capability scaling as a new paradigm for multimodal comprehension.
AutoMem: Automated framework trains LLMs to manage memory as a learnable cognitive skill
AutoMem is a new framework that treats memory management in LLMs as a trainable skill, using two optimization loops: one that iteratively revises memory structure via trajectory review by a strong LLM, and one that distills good memory decisions into direct training signal for the agent model. Evaluated on three long-horizon procedurally generated games (Crafter, MiniHack, NetHack), optimizing memory alone yielded 2x-4x performance improvements, bringing a 32B open-weight model competitive with frontier systems like Claude Opus 4.5 and Gemini 3.1 Pro Thinking. The work draws on cognitive science concepts of metamemory and demonstrates that memory management is an independently learnable, high-leverage capability for long-horizon agentic tasks.