CONTRAMEM: Learning Self-Evolving Procedural Memory from Contrasting Multi-Model Trajectories

• URL: http://arxiv.org/abs/2608.22533v1
• Date: Sun, 23 Aug 2026 18:11:26 GMT
• Title: CONTRAMEM: Learning Self-Evolving Procedural Memory from Contrasting Multi-Model Trajectories
• Abstract summary: CONTRAMEM is a source-flexible, training-free framework for self-evolving procedural memory.<n>On held-out GAIA2/ARE computer-use tasks, CONTRAMEM more than doubles the success rate.
• Score: 13.649910962870896
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Autonomous computer-use agents are increasingly applied to long-horizon tasks requiring coordinated application calls, persistent state tracking, and verifier-sensitive writes, yet they remain prone to procedural failures: misreading application state, tool semantics, or task progress. Procedural memory promises more consistent decisions and less redundant exploration, but constructing high-quality memory without model training remains challenging. We introduce CONTRAMEM, a source-flexible, training-free framework for self-evolving procedural memory that treats same-task outcome variation as supervision: differences in correctness, efficiency, recovery, and failure modes expose outcome-relevant procedural distinctions, distilled into a compact bank of app-level Function Cards and task-level Skill Cards that evolves through localized curation rather than append-only accumulation or whole-bank rewriting. On held-out GAIA2/ARE computer-use tasks, CONTRAMEM more than doubles the success rate across the three source-model targets (26.2% to 55.3%), with consistent per-model gains (GPT-5.5: 27.5 to 61.0; Claude Sonnet 4.6: 28.0 to 52.5; DeepSeek V4 Pro: 23.0 to 52.5). The same bank transfers unchanged to the unseen Qwen3.7 Plus (18.5 to 35.5), indicating transferable procedural knowledge rather than model-specific behavior. The same construction carries over unchanged to AppWorld, beating both no memory and its own single-source self-memory variant for all three mid-tier agents on both public test splits. Under a matched trajectory budget, heterogeneous multi-model trajectories yield stronger memory than self- or same-model multi-rollout memory: the margin comes from contrastive behavioral diversity, not stronger source agents or more sampling.

Related papers

• What Makes Agent Memory Useful for Reliable Unanswerable Question Handling? [51.58639607746329]
  Memory is widely used in agent systems, but its role in reliable UAQ handling remains unclear.<n>We present a systematic study of agent memory for UAQ handling under a unified agentic RAG framework.<n>We find that memory can improve UAQ performance in some settings, but such gains are selective rather than universal and remain fragile under dataset shift.
  arXiv  Detail & Related papers  (2026-08-28T05:00:28Z)
• Same Model, Different Harness: Different Coding-Agent Results [0.0]
  We compare two configurations of the same harness on three coding benchmarks.<n>The control supplies the full conversation in time order, while the treatment keeps the same record but mechanically shortens older tool results.<n>Under tight context, the treatment raises mean per-task fail-to-pass fraction (F2PF) in all three pressure comparisons.
  arXiv  Detail & Related papers  (2026-08-26T11:55:44Z)
• Recursive Experiential-Working Memory Evolution for Long-Horizon Agent Harnesses [83.32611586724165]
  Recuris is a Experiential-Working Memory architecture for long-horizon agent harnesses.<n>Working Memory tracks task progress and guides skill selection from Experiential Memory.<n>Across four long-horizon benchmarks and ten models, Recuris improves task success in 35 of the 37 completed model-benchmark pairs.
  arXiv  Detail & Related papers  (2026-08-25T17:56:35Z)
• TRACE: A Self-Evolving Skill Bank for Consistent, Limit-Aware LLM Agents [41.82249291285459]
  TRACE iteratively improves a skill-based agent's behavioral knowledge without modifying model weights.<n>On GPT-5.5, TRACE improves consistency (Pass3) by 34.6 points, from 59.9% to 94.5%, while shrinking the gap between potential and reliable performance to just 4.0 points.<n>These results show that TRACE converts high model potential into stable, consistent performance gain.
  arXiv  Detail & Related papers  (2026-08-24T04:36:03Z)
• Agent Memory Distillation: Empowering Small LLM Agents with Hierarchical Teacher Memory [49.6908961629532]
  Agent Memory Distillation (AMD) is a training-free framework that transfers structured knowledge from a large teacher agent to a small student agent through hierarchical memory.<n>AMD constructs three complementary memory types from successful teacher trajectories: memory encodes task-level strategies, Subtask memory provides concrete behavioral examples, and Function memory captures per-function calling conventions.<n>We evaluate AMD on three tool-use benchmarks using four student models (4B-8B parameters) with GPT-5-mini as the teacher, achieving average accuracy gains of 27.2%p, 11.2%p, and 3.4%p on AppWorld, BFCL V
  arXiv  Detail & Related papers  (2026-08-07T12:43:00Z)
• Self-Play Meets Skill Evolution: Self-Evolving Search Agents that Pose, Solve, and Remember [26.754732469136385]
  Self-play agents can generate training problems without questions from target benchmarks, but their curricula lack persistent state.<n>We introduce textbfSESA (Self-Evolving Skill-Augmented Agent), which makes procedural memory an evolving state of tool-augmented search self-play.
  arXiv  Detail & Related papers  (2026-07-31T14:32:52Z)
• AutoMem: Automated Learning of Memory as a Cognitive Skill [63.07128029793901]
  We show that memory expertise is a learned skill.<n>We introduce AutoMem, a framework that automates the two axes of memory management.<n>Our results show that memory management is an independently learnable skill.
  arXiv  Detail & Related papers  (2026-07-01T17:57:03Z)
• SaliMory: Orchestrating Cognitive Memory for Conversational Agents [22.813290545851743]
  SALIMORY is a framework that trains a single language model to manage a cognitively-structured memory-spanning user facts, preferences, and working memory.<n>It cuts memory-attributed failures by one-third, outperforms the state-of-the-art by over 10% in end-to-end accuracy, and more than doubles the Good Personalization rate.
  arXiv  Detail & Related papers  (2026-06-02T18:31:50Z)
• 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)
• Memp: Exploring Agent Procedural Memory [72.41472703974935]
  Large Language Models (LLMs) based agents excel at diverse tasks, yet they suffer from brittle procedural memory that is manually engineered or entangled in static parameters.<n>We propose Memp that distills past agent trajectories into both fine-grained, step-by-step instructions and higher-level, script-like abstractions.<n>We show that as the memory repository is refined, agents achieve steadily higher success rates and greater efficiency on analogous tasks.
  arXiv  Detail & Related papers  (2025-08-08T16:20:56Z)
• Sable: a Performant, Efficient and Scalable Sequence Model for MARL [3.676220008456203]
  We introduce Sable, a performant, memory-efficient, and scalable sequence modeling approach to MARL.<n> Sable works by adapting the retention mechanism in Retentive Networks.<n>We demonstrate how Sable is able to significantly outperform existing state-of-the-art methods in a large number of diverse tasks.
  arXiv  Detail & Related papers  (2024-10-02T16:15:26Z)

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.