When to Forget: A Memory Governance Primitive

• URL: http://arxiv.org/abs/2604.12007v1
• Date: Mon, 13 Apr 2026 19:54:14 GMT
• Title: When to Forget: A Memory Governance Primitive
• Abstract summary: Memory Worth is a two-counter per-memory signal that tracks how often a memory co-occurs with successful versus failed outcomes.<n>We prove that MW converges almost surely to the conditional success probability p+(m) = Pr[y_t = +1 | m in M_t] under a stationary retrieval regime.<n>We argue this is still a useful operational signal for memory governance, and we validate it empirically in a controlled synthetic environment.
• Score: 0.0
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Agent memory systems accumulate experience but currently lack a principled operational metric for memory quality governance -- deciding which memories to trust, suppress, or deprecate as the agent's task distribution shifts. Write-time importance scores are static; dynamic management systems use LLM judgment or structural heuristics rather than outcome feedback. This paper proposes Memory Worth (MW): a two-counter per-memory signal that tracks how often a memory co-occurs with successful versus failed outcomes, providing a lightweight, theoretically grounded foundation for staleness detection, retrieval suppression, and deprecation decisions. We prove that MW converges almost surely to the conditional success probability p+(m) = Pr[y_t = +1 | m in M_t] -- the probability of task success given that memory m is retrieved -- under a stationary retrieval regime with a minimum exploration condition. Importantly, p+(m) is an associational quantity, not a causal one: it measures outcome co-occurrence rather than causal contribution. We argue this is still a useful operational signal for memory governance, and we validate it empirically in a controlled synthetic environment where ground-truth utility is known: after 10,000 episodes, the Spearman rank-correlation between Memory Worth and true utilities reaches rho = 0.89 +/- 0.02 across 20 independent seeds, compared to rho = 0.00 for systems that never update their assessments. A retrieval-realistic micro-experiment with real text and neural embedding retrieval (all-MiniLM-L6-v2) further shows stale memories crossing the low-value threshold (MW = 0.17) while specialist memories remain high-value (MW = 0.77) across 3,000 episodes. The estimator requires only two scalar counters per memory unit and can be added to architectures that already log retrievals and episode outcomes.

Related papers

• Caching for the Future: Scrub Jay Episodic Memory Principles for Agent Memory Systems [5.321970297768025]
  Existing memory architectures treat all memories as equally persistent and systematically contaminate retrieved context with outdated facts.<n>We show that per-memory, type-conditioned temporal decay, can be operationalized as an auto-classified coefficient $_i$ in an external LLM-agent memory store.<n>We introduce the Temporal Generalization Test (TGT), a benchmark with held-out retention intervals and a Generalization Gap (GenGap) metric.
  arXiv  Detail & Related papers  (2026-08-05T12:12:44Z)
• Beyond Memory Leaderboards: Evaluating Scientific Memory as Budgeted Context Restoration [0.05097809301149341]
  We introduce two full-text scientific-memory benchmarks, Public AI Memory (PAIM) and Public Transformers (PTr)<n>We evaluate eight memory/retrieval systems, including our own proposed Theoria, plus a no-retrieval baseline.<n>Our results show that memory leaderboards are not interpretable without the full protocol.
  arXiv  Detail & Related papers  (2026-07-18T15:09:58Z)
• Memory as a Controlled Process: Learned Adaptive Memory Management for LLM Agents [72.3752981234916]
  Large Language Model (LLM) agents increasingly rely on external memory systems to accumulate experience across tasks.<n>We present Memory as a Controlled Process (MemCon), a framework that models memory operations as a Markov Decision Process.<n>MemCon consistently outperforms multiple memory baselines by up to 15.2 points in task success while reducing token consumption by 5-20%.
  arXiv  Detail & Related papers  (2026-07-15T08:32:40Z)
• MemOps: Benchmarking Lifecycle Memory Operations in Long-Horizon Conversations [50.24315431387575]
  We introduce MemOps, a benchmark that reformulates conversational memory as a sequence of lifecycle operations.<n>MemOps represents each memory event with a structured trace specifying its trigger, target, scope, state transition, and supporting evidence.<n>Across long-context, retrieval-based, parametric and managed-memory systems, MemOps disentangles failure modes that final-answer accuracy alone conceals.
  arXiv  Detail & Related papers  (2026-07-14T15:33:44Z)
• The Compliance Trap: Diagnosing How AI Agents Consume Conflicting Memory [9.041305193333434]
  Memory is becoming a core component of long-horizon AI agents, allowing agents to reuse past experience when operating web browsers, software tools, and other interactive environments.<n>Existing work mostly treats memory as a supply problem, asking what experience to write, how to store it, and which entry to retrieve for the next task.<n>Yet we still lack a clear account of how models consume retrieved memory across a multi-step action trajectory.
  arXiv  Detail & Related papers  (2026-07-12T07:05:31Z)
• MEMPROBE: Probing Long-Term Agent Memory via Hidden User-State Recovery [58.64329475683699]
  We argue that long-term memory should be evaluated as an auditable post-interaction artifact.<n>We instantiate this view in MEMPROBE, a benchmark in which a memory-equipped agent assists simulated users.<n>We find that successful assistance and recoverable memory behave as distinct capabilities.
  arXiv  Detail & Related papers  (2026-06-23T13:52:46Z)
• Learning What to Remember: A Cognitively Grounded Multi-Factor Value Model for Agentic Memory [1.2127875744950842]
  Long-running LLM agents accumulate interaction histories far larger than any context window.<n>Production systems answer with semantic similarity or recency.<n>We propose a multi-factor memory function V(m)=sum_i w_i f_i(m) over seven interpretable factors.
  arXiv  Detail & Related papers  (2026-06-11T06:17:31Z)
• Meta-Cognitive Memory Policy Optimization for Long-Horizon LLM Agents [13.846550145896314]
  Memory-augmented LLM agents tackle complex long-horizon tasks by summarizing interaction trajectories into compact memory.<n>Existing approaches typically train these memory policies using outcome-based reinforcement learning, failing to localize where intermediate memory quality degrades.<n>We argue that memory optimization should focus not merely on trajectory-level success, but on the clarity of the belief induced by intermediate summaries.
  arXiv  Detail & Related papers  (2026-05-28T16:17:19Z)
• WorldMemArena: Evaluating Multimodal Agent Memory Through Action-World Interaction [72.1620416874118]
  Multimodal large language models are increasingly deployed as long-horizon agents.<n>Existing benchmarks measure recall over static dialogue, collapse memory into a single end-of-task accuracy, and reduce visual observations to captions.<n>We formulate multimodal agent memory as an Action-World Interaction Loop with an observable four-stage lifecycle.
  arXiv  Detail & Related papers  (2026-05-28T04:27:20Z)
• MemGym: a Long-Horizon Memory Environment for LLM Agents [69.79226770543049]
  We present MemGym, a benchmark for agentic memory that unifies existing agent gyms and in-house memory-grounded pipelines.<n>MemGym reports memory-isolated scores that decouple memory performance from reasoning, retrieval, and tool-use ability.<n>Our synthetic pipelines for MEMGYM-CODEQA and MEMGYM-DR are length-controllable, ablation-verified at every stage, and tightly aligned with downstream scenarios.
  arXiv  Detail & Related papers  (2026-05-20T07:25:33Z)
• Is One Score Enough? Rethinking the Evaluation of Sequentially Evolving LLM Memory [50.857546269660276]
  We introduce SeqMem-Eval, a diagnostic evaluation framework for sequentially evolving large language models (LLMs) memory.<n>Rather than focusing only on final performance, SeqMem-Eval evaluates how memory states evolve, generalize, consolidate experience, and retain useful information during sequential inference.
  arXiv  Detail & Related papers  (2026-05-14T20:15:22Z)
• ImplicitMemBench: Measuring Unconscious Behavioral Adaptation in Large Language Models [60.14219417402433]
  Existing memory benchmarks for LLM agents evaluate explicit recall of facts, yet overlook implicit memory where experience becomes automated behavior without conscious retrieval.<n>We introduce ImplicitMemBench, the first systematic benchmark evaluating implicit memory through three constructs drawn from standard cognitive-science accounts of non-declarative memory.<n>Our 300-item suite employs a unified Learning/Priming-Interfere-Test protocol with first-attempt scoring.
  arXiv  Detail & Related papers  (2026-04-09T10:26:32Z)
• MemMachine: A Ground-Truth-Preserving Memory System for Personalized AI Agents [2.541923091180284]
  Large Language Model (LLM) agents require persistent memory to maintain personalization, factual continuity, and long-horizon reasoning.<n>We present MemMachine, an open-source memory system that integrates short-term, long-term episodic, and profile memory.<n>MemMachine uses contextualized retrieval that expands nucleus matches with surrounding context, improving recall when relevant evidence spans multiple dialogue turns.
  arXiv  Detail & Related papers  (2026-04-06T16:57:06Z)
• From Verbatim to Gist: Distilling Pyramidal Multimodal Memory via Semantic Information Bottleneck for Long-Horizon Video Agents [78.30630000529133]
  We propose MM-Mem, a pyramidal multimodal memory architecture grounded in Fuzzy-Trace Theory.<n> MM-Mem memory structures hierarchically into a Sensory Buffer, Episodic Stream, and Symbolic.<n>Experiments confirm the effectiveness of MM-Mem on both offline and streaming tasks.
  arXiv  Detail & Related papers  (2026-03-02T05:12:45Z)
• KEEP: A KV-Cache-Centric Memory Management System for Efficient Embodied Planning [8.216400469571084]
  We propose KEEP, a KV-cache-centric memory management system for efficient embodied planning.<n>KEEP features 3 key innovations: (1) a Static-Dynamic Memory Construction algorithm that reduces KV cache recomputation by mixed-granularity memory group; (2) a Multi-hop Memory Re-computation algorithm that dynamically identifies important cross-attention among different memory groups; and (3) a Layer-balanced Memory Loading that eliminates unbalanced KV cache loading and cross-attention across different layers.
  arXiv  Detail & Related papers  (2026-02-27T01:48:07Z)
• MemCtrl: Using MLLMs as Active Memory Controllers on Embodied Agents [53.44122827359892]
  We propose MemCtrl, a framework that uses Multimodal Large Language Models (MLLMs) for pruning memory online.<n>-augmented MLLMs show an improvement of around 16% on average, with over 20% on specific instruction subsets.
  arXiv  Detail & Related papers  (2026-01-28T18:31:17Z)
• Entropic Efficiency of Bayesian Inference Protocols [0.8594140167290097]
  Inference is a versatile tool that underlies scientific discovery, machine learning, and everyday decision-making.<n>We define an inferential efficiency as the ratio of information gain to cumulative memory erasure cost.<n>We benchmark two limiting measurement paradigms: sequential, in which the same memory is exploited iteratively, and parallel, in which many memories are exploited simultaneously.
  arXiv  Detail & Related papers  (2026-01-24T03:43:38Z)
• MemRec: Collaborative Memory-Augmented Agentic Recommender System [57.548438733740504]
  We propose MemRec, a framework that architecturally decouples reasoning from memory management.<n>MemRec introduces a dedicated LM_Mem to manage a dynamic collaborative memory graph.<n>It achieves state-of-the-art performance on four benchmarks.
  arXiv  Detail & Related papers  (2026-01-13T18:51:16Z)
• Mnemosyne: An Unsupervised, Human-Inspired Long-Term Memory Architecture for Edge-Based LLMs [1.2582867366903179]
  We introduce Mnemosyne, an unsupervised, human-inspired long-term memory architecture for edge-based large language models (LLM)<n>Our approach uses graph-structured storage, modular substance and redundancy filters, memory committing and pruning mechanisms, and probabilistic recall with temporal decay and refresh processes modeled after human memory.<n>Mnemosyne is designed for use in longitudinal healthcare assistants, where repetitive and semantically similar but temporally distinct conversations are limited by naive retrieval.
  arXiv  Detail & Related papers  (2025-10-07T03:32:53Z)
• Memorize or Generalize? Evaluating LLM Code Generation with Code Rewriting [54.48306552577881]
  We argue that large language models (LLMs) are mostly doing memorization (i.e., replicating or reusing large parts of their training data) versus generalization.<n>Existing evaluations largely proxy neglecting surface/structural similarity, thereby conflating benign reuse of repeated code with harmful recall and memorization task correctness.<n>We propose Memorization Risk Index (MRI), a normalized score that combines two signals: (i) how similar the model's answer for the rewritten task is to the original ground-truth solution, and (ii) how much performance drops from the original task to its rewritten counterpart.
  arXiv  Detail & Related papers  (2025-03-04T05:39:24Z)
• MADial-Bench: Towards Real-world Evaluation of Memory-Augmented Dialogue Generation [15.64077949677469]
  We present a novel Memory-Augmented Dialogue Benchmark (MADail-Bench) to evaluate the effectiveness of memory-augmented dialogue systems (MADS) The benchmark assesses two tasks separately: memory retrieval and memory recognition with the incorporation of both passive and proactive memory recall data. Results from cutting-edge embedding models and large language models on this benchmark indicate the potential for further advancement.
  arXiv  Detail & Related papers  (2024-09-23T17:38:41Z)
• RMM: Reinforced Memory Management for Class-Incremental Learning [102.20140790771265]
  Class-Incremental Learning (CIL) trains classifiers under a strict memory budget. Existing methods use a static and ad hoc strategy for memory allocation, which is often sub-optimal. We propose a dynamic memory management strategy that is optimized for the incremental phases and different object classes.
  arXiv  Detail & Related papers  (2023-01-14T00:07:47Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.