Memory Depth, Not Memory Access: Selective Parametric Consolidation for Long-Running Language Agents

• URL: http://arxiv.org/abs/2606.26806v1
• Date: Thu, 25 Jun 2026 09:42:56 GMT
• Title: Memory Depth, Not Memory Access: Selective Parametric Consolidation for Long-Running Language Agents
• Abstract summary: We study memory depth: durable goal-conditioned tendencies written into a small parametric store.<n>We introduce the loop-drift protocol, a controlled stress test in which the retrieval index remains intact while working context is unloaded.<n>We find that selective consolidation factorizes into two controllable dimensions: selection and actuation.
• Score: 0.0
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Long-running language agents need more than memory access. Retrieval systems can fetch past facts at query time, but they do not decide which experiences should continue to shape behavior after the working context is unloaded. We study this separate problem as memory depth: durable goal-conditioned tendencies written into a small parametric store. We introduce the loop-drift protocol, a controlled stress test in which the retrieval index remains intact while working context is unloaded and goal-conditioned behavior must persist under long-loop interference. We evaluate EVAF, a surprise- and valence-gated LoRA consolidation mechanism. Across GPT-2 and TinyLlama, retrieval is strongest on shallow factual recall (short-fact accuracy 0.956--0.973), while EVAF is strongest on goal persistence and post-unload recovery (0.812--0.904) with only 2--3 parametric writes per 200 events. Mechanism controls show that selective consolidation factorizes into two controllable dimensions: selection and actuation. Matched random gates isolate selection beyond sparse writing; fixed-inner controls across GPT-2, TinyLlama, and Mistral-7B show that inner-loop write strength is model-dependent; and a Mistral-7B matched-gate inversion reveals asymmetric selection-actuation coupling under miscalibrated actuation. Public Memora event streams serve as an external diagnostic, exposing stale-memory invalidation as an unresolved boundary. Within this probe, selective parametric consolidation supplies memory depth distinct from and complementary to retrieval access.

Related papers

• Hindsight Memory-PRM: Supervising Memory Management with Auditable Hindsight Credit [48.811961095386]
  Hindsight Memory-PRM exploits this audit trail twice.<n> offline to train an operation-conditioned memory-utility critic.<n>Online, where retrievals, citations, and one controlled deletion-and-reanswer per probe settle an intervention-calibrated entry-level presence credit.
  arXiv  Detail & Related papers  (2026-08-30T06:53:13Z)
• 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)
• EVAF: A Test-Retest Protocol for Selective Parametric Consolidation [0.0]
  We introduce EVAF, an Echo-Valence Attractor Field mechanism for gated LoRA consolidation.<n>Across GPT-2 and TinyLlama, EVAF preferentially consolidates high-valence, high-surprise experiences.<n>Test-retest measurements show stronger post-interference behavioral persistence than frozen, retrieval-only, and ungated continual-update baselines.
  arXiv  Detail & Related papers  (2026-06-29T07:53:50Z)
• ConvMemory v3: A Validity Context Layer for Conversational Memory via Target-Conditioned Relation Verification [0.0]
  ConvMemory v3 adds a validity context layer that detects and surfaces this update evidence.<n>The core mechanism is a dual-evidence gate that conditions a relation judgment on the specific target proposition.<n>On a synthetic multi-hop validity benchmark the gate reaches 90.12% +/- 1.73 accuracy.
  arXiv  Detail & Related papers  (2026-06-25T08:36:43Z)
• 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)
• SAM: State-Adaptive Memory for Long-Horizon Reasoning Agent [51.274445160155864]
  Long-horizon agentic reasoning requires large language models to act over long interaction histories containing thoughts, tool calls, observations, and partial conclusions.<n>Existing approaches address this difficulty by truncating the interaction history, compressing it into shorter surrogates, or retrieving selected parts of it for reuse.<n>We propose State-Adaptive Memory, a standalone framework that consolidates ongoing interaction into compact memory cues while preserving raw trajectory pages for intent-driven recall.
  arXiv  Detail & Related papers  (2026-05-23T08:37:16Z)
• GoTTA be Diverse: Rethinking Memory Policies for Test-Time Adaptation [54.70334474895361]
  Test-time adaptation (TTA) enables a pre-trained model to adapt online to an unlabeled test stream under distribution shift.<n>While most TTA research focuses on the adaptation objective, practical streams also depend critically on the memory used to select which test samples drive adaptation.<n>We introduce Guided Observational Test-Time Adaptation (GOTTA), a family of diversity-aware memory policies that combine class-balanced allocation with feature-space diversity.
  arXiv  Detail & Related papers  (2026-05-19T14:18:46Z)
• Temporal Attention for Adaptive Control of Euler-Lagrange Systems with Unobservable Memory [1.2891210250935148]
  The paper proposes a meta-control architecture in which the gains of a computed-torque controller are generated by a self-attention block.<n>The number of attention heads is selected before policy training through a surrogate analysis of the memory-state gradient.<n>The approach is tested on a 2-DOF manipulator with nonlinear friction and variable payload.
  arXiv  Detail & Related papers  (2026-05-07T19:25:18Z)
• Evoking User Memory: Personalizing LLM via Recollection-Familiarity Adaptive Retrieval [59.295767860331004]
  RF-Mem is a familiarity uncertainty-guided dual-path memory retriever.<n>It embeds human-like dual-process recognition into the retriever.<n>It consistently outperforms one-shot retrieval and full-context reasoning under fixed budget and latency constraints.
  arXiv  Detail & Related papers  (2026-03-10T06:31:44Z)
• TS-Memory: Plug-and-Play Memory for Time Series Foundation Models [63.21390142212087]
  Time Series Foundation Models (TSFMs) achieve strong zero-shot forecasting through large-scale pre-training.<n>Existing solutions face a trade-off: Parametric Adaptation can cause catastrophic forgetting, while Non-Parametric Retrieval improves forecasts but incurs high latency due to datastore search.<n>We propose Parametric Memory Distillation and implement it as TS-Memory, a lightweight memory adapter that augments frozen TSFMs.
  arXiv  Detail & Related papers  (2026-02-12T04:16:19Z)
• Agentic Unlearning: When LLM Agent Meets Machine Unlearning [45.60284810359578]
  We present Synchronized Backflow Unlearning (SBU), a framework that unlearns jointly across parameter and memory pathways.<n>SBU reduces traces of targeted private information across both pathways with limited degradation on retained data.
  arXiv  Detail & Related papers  (2026-02-06T07:13:13Z)
• Memory in Large Language Models: Mechanisms, Evaluation and Evolution [8.158439933515131]
  We propose a four-part taxonomy (parametric, contextual, external, procedural/episodic) and a memory quadruple (location, persistence, write/access path, controllability)<n>For updating and forgetting, we present DMM Gov: coordinating DAPT/TAPT, PEFT, model editing (ROME, MEND, MEMIT, SERAC), and RAG to form an auditable loop.<n>This yields a reproducible, comparable, and governable coordinate system for research and deployment.
  arXiv  Detail & Related papers  (2025-09-23T10:06:58Z)
• In Prospect and Retrospect: Reflective Memory Management for Long-term Personalized Dialogue Agents [70.12342024019044]
  Large Language Models (LLMs) have made significant progress in open-ended dialogue, yet their inability to retain and retrieve relevant information limits their effectiveness.<n>We propose Reflective Memory Management (RMM), a novel mechanism for long-term dialogue agents, integrating forward- and backward-looking reflections.<n>RMM shows more than 10% accuracy improvement over the baseline without memory management on the LongMemEval dataset.
  arXiv  Detail & Related papers  (2025-03-11T04:15:52Z)

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.