# AutoMem: A Text-Gradient Recursive Self-Improvement Framework for Automated Memory Architectures Search

arXiv: https://arxiv.org/abs/2608.14621 · HTML: https://mlsanity.com/paper/2608.14621/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Agent memory systems, one of 16 directions in LLM agents and tool use.
- Builds on 8 of its 17 references that are in the map.
- First posted on arXiv 14 Jul 2026.

## Placed in

- LLM agents and tool use / [Agent memory systems](https://mlsanity.com/agents/agent-memory-systems/)
- Synthetic data and self-improvement for LLMs / [Self-evolving LLM agents](https://mlsanity.com/synth/self-evolving-llm-agents/)

## Builds on (most cited first, 8 of 8)

- [MemoryBank: Enhancing Large Language Models with Long-Term Memory](https://mlsanity.com/paper/2305.10250/) (2023)
- [Precise Zero-Shot Dense Retrieval without Relevance Labels](https://mlsanity.com/paper/2212.10496/) (2022)
- [Darwin Godel Machine: Open-Ended Evolution of Self-Improving Agents](https://mlsanity.com/paper/2505.22954/) (2025)
- [Which Agent Causes Task Failures and When? On Automated Failure Attribution of LLM Multi-Agent Systems](https://mlsanity.com/paper/2505.00212/) (2025)
- [Memento: Fine-tuning LLM Agents without Fine-tuning LLMs](https://mlsanity.com/paper/2508.16153/) (2025)
- [Memory for Autonomous LLM Agents:Mechanisms, Evaluation, and Emerging Frontiers](https://mlsanity.com/paper/2603.07670/) (2026)
- [Building Self-Evolving Agents via Experience-Driven Lifelong Learning: A Framework and Benchmark](https://mlsanity.com/paper/2508.19005/) (2025)
- [Flash-Searcher: Fast and Effective Web Agents via DAG-Based Parallel Execution](https://mlsanity.com/paper/2509.25301/) (2025)

## Built on by (most cited first, 2 of 2)

- [Vestrum: Improving Agent Harnesses by Adapting Their Verification, Structure and Memory](https://mlsanity.com/paper/2609.33822/) (2026)
- [MemCodex: Self-Programming Hierarchical Memory for Language Agents](https://mlsanity.com/paper/2609.39765/) (2026)

## Abstract (arXiv)

Long-term memory is increasingly central to LLM agents, yet memory design remains a highly coupled architecture problem: what to encode, how to store it, how to retrieve it, and how to manage it can vary substantially across tasks and backbone models. We construct a discrete search space with 5 encoders, 5 stores, 6 retrievers, and 4 managers, and show that no single memory architecture consistently dominates: different tasks favor different module combinations, leading to substantial performance gaps. Motivated by this, we propose textscAutoMem, a text-gradient recursive self-improvement framework for task-adaptive memory architecture search. textscAutoMem optimizes over the factored space through two components: Experience-Guided Architecture Search, which proposes candidate architectures from historical search trajectories and accumulated reflections, and Failure-Guided Module Diagnosis, which localizes memory-related failures to specific modules and converts them into targeted textual feedback. Experiments on GAIA, WebWalkerQA, and xBench-DeepSearch across two LLM backbones show that textscAutoMem consistently discovers task-adaptive memory architectures that outperform the strongest human-designed memory baselines, improving accuracy by 2.8 points on average across six benchmark-backbone settings. Further analysis shows that textscAutoMem achieves a favorable accuracy-efficiency trade-off, reducing token cost by 14.3% over the strongest accuracy baselines under Qwen3.5-122B-A10B, while also finding stronger architectures than substantially larger random searches within only a few guided iterations.