GROUNDING

Agent Memory

6 min read

Agent Memory is the set of mechanisms that lets AI agents persist useful context across turns, sessions, tools, and tasks. GROUNDING tracks memory stores, compression, retrieval boundaries, and failure modes for long-running agent workflows.

Topic: Agents Related: Agents Context Engineering RAG

Recent Updates

  • 2026-06-05: EMBER: Efficient Memory via Budgeted Evidence Retention for Long-Horizon Agents (cs.CL updates on arXiv.org) · arxiv.org — EMBER-14B
  • 2026-06-05: Generic Triple-Latent Compression with Gated Associative Retrieval (cs.CL updates on arXiv.org) · arxiv.org
  • 2026-06-05: A Motivational Architecture for Conversational AGI (cs.AI updates on arXiv.org) · arxiv.org
  • 2026-06-05: Harnessing Generalist Agents for Contextualized Time Series (cs.AI updates on arXiv.org) · arxiv.org
  • 2026-06-05: Bootstrapping Semantic Layer from Execution for Text-to-SQL (cs.CL updates on arXiv.org) · arxiv.org
  • 2026-06-05: When Should Memory Stay Silent: Measuring Memory-Use Boundaries in Memory-Augmented Conversational Agents (cs.AI updates on arXiv.org) · arxiv.org — OpenAI Anthropic DeepSeek Alibaba Cloud GPT-5.4-mini Claude Sonnet 4.6 DeepSeek V4 Flash Qwen3.5-9B
  • 2026-06-05: LANTERN: A Lightweight Memory Layer for Long-Context Conversations (cs.CL updates on arXiv.org) · arxiv.org
  • 2026-06-05: ArcANE: Do Role-Playing Language Agents Stay in Character at the Right Time? (cs.CL updates on arXiv.org) · arxiv.org — arXiv Hugging Face CatalyzeX DagsHub Gotit.pub ScienceCast ArcANE-8B ArcANE-32B
  • 2026-06-05: QueryAgent-R1: Bridging Query Generation and Product Retrieval for E-Commerce Query Recommendation (cs.CL updates on arXiv.org) · arxiv.org
  • 2026-06-05: SentinelBench: A Benchmark for Long-Running Monitoring Agents (cs.AI updates on arXiv.org) · arxiv.org — arXiv Amanda Swearngin
  • 2026-06-05: Rethinking LoRA Memory Through the Lens of KV Cache Compression (cs.CL updates on arXiv.org) · arxiv.org
  • 2026-06-05: LatentSkill: Moving Agent Procedures from Context to Weights (cs.CL updates on arXiv.org) · arxiv.org
  • 2026-06-05: Self-supervised User Profile Generation for Personalization (BUMP) (cs.CL updates on arXiv.org) · arxiv.org
  • 2026-06-05: Continual Learning Bench: Evaluating Frontier AI Systems in Real-World Stateful Environments (cs.AI updates on arXiv.org) · arxiv.org
  • 2026-06-05: Fix the Mind, Not the Move: Interpretable AI Assistance via Knowledge-Gap Localization (cs.AI updates on arXiv.org) · arxiv.org — arXiv
  • 2026-06-05: EpiEvolve: Self-Evolving Agents for Streaming Pandemic Forecasting (cs.AI updates on arXiv.org) · arxiv.org — CDC
  • 2026-06-05: How LLM Agents Work: Architecture, Planning, and Tools (Искусственный интеллект – AI, ANN и иные формы искусственного разума) · habr.com — GPT-5
  • 2026-06-05: Analysis of the Odysseus Self-Hosted AI Workspace (Искусственный интеллект – AI, ANN и иные формы искусственного разума) · habr.com — Alibaba OpenAI Anthropic PewDiePie
  • 2026-06-05: Automnemomorph: Philosophical Challenges of Absolute Memory Control in Agents (Искусственный интеллект – AI, ANN и иные формы искусственного разума) · habr.com
  • 2026-06-06: Modular Architecture as an Alternative to Long Context in AI-Assisted Coding (Искусственный интеллект – AI, ANN и иные формы искусственного разума) · habr.com
  • 2026-06-06: CL-Bench: A Benchmark for Continual Learning in AI Agents (DAIR.AI) · arxiv.org — DAIR.AI omarsar0
  • 2026-06-06: Five Labs, Five Minds: Engineering Lessons from Multi-Model Agent Systems (Hugging Face - Blog) · huggingface.co — Hugging Face OpenAI OpenBMB NVIDIA gpt-oss-20b MiniCPM3-4B Nemotron-Mini-4B Qwen 0.5B
  • 2026-06-06: Universal Memory Protocol – a shared format for agent memory (Hacker News) · universalmemoryprotocol.io — PostgreSQL Redis SQLite Pinecone Weaviate Qdrant Anthropic OpenAI
  • 2026-06-06: Hermes-Agent: A Self-Improving Agent from Nous Research (GitHub AI Ranking Changes (Top 10)) · github.com — Nous Research OpenRouter NovitaAI NVIDIA Xiaomi z.ai GLM Kimi Moonshot MiniMax Hugging Face OpenAI Firecrawl fal Discord Slack WhatsApp Signal Nemotron
  • 2026-06-06: Hermes-agent by Nous Research (GitHub AI Ranking Changes (Top 10)) · github.com — Nous Research OpenAI OpenRouter NovitaAI NVIDIA Xiaomi z.ai Moonshot MiniMax Hugging Face fal Firecrawl Nemotron
  • 2026-06-06: Google Introduces Agentic RAG Framework for Gemini Enterprise Agent Platform (The latest research from Google) · research.google — Google Google Research Google Cloud Cyrus Rashtchian Da-Cheng Juan Gemini
  • 2026-06-07: Context Sculpting (Hacker News) · perceptiontheory.bearblog.dev — Anthropic OpenAI Viv Opus Sonnet Haiku Claude
  • 2026-06-07: Claude-Mem: Persistent Context and Memory for Agentic Workflows (GitHub AI Ranking Changes (Top 10)) · github.com — Anthropic Google
  • 2026-06-08: MemDreamer: Decoupling Perception and Reasoning for Long Video Understanding via Hierarchical Graph Memory and Agentic Retrieval Mechanism (cs.CL updates on arXiv.org) · arxiv.org
  • 2026-06-08: Tree-of-Experience: A Structured Experience-Management Method for Self-Evolving Agents (cs.CL updates on arXiv.org) · arxiv.org
  • 2026-06-08: Autonomous Heterogeneous Catalyst Discovery with CatDT (cs.AI updates on arXiv.org) · arxiv.org
  • 2026-06-08: MExam: Benchmarking Multimodal Memory for Realistic User-Agent Interactions (cs.CL updates on arXiv.org) · arxiv.org
  • 2026-06-08: AdMem: A Unified Memory Framework for Task-Solving Agents (cs.AI updates on arXiv.org) · arxiv.org
  • 2026-06-09: Personalization Meets Safety: Mechanisms, Risks, and Mitigations in Personalized LLMs (cs.AI updates on arXiv.org) · arxiv.org
  • 2026-06-09: Co-Evolving Skill Generation and Policy Optimization (cs.CL updates on arXiv.org) · arxiv.org
  • 2026-06-09: Bayesian-Agent: Posterior-Guided Skill Evolution for LLM Agent Harnesses (cs.CL updates on arXiv.org) · arxiv.org — DeepSeek DeepSeek V4 Flash
  • 2026-06-09: ConMem: Structured Memory-Guided Adaptation in Training-Free Multi-Agent Systems (cs.AI updates on arXiv.org) · arxiv.org
  • 2026-06-09: Multimodal LLM Agents Fail to Develop Partner-Specific Conventions in Collaborative Tasks (cs.CL updates on arXiv.org) · arxiv.org — Po-Ya Angela Wang
  • 2026-06-09: Customer-Agent: Overcoming Context Limitations in Ultra-Long Shopping Trajectories via Tool-Augmented Agents and RLVR (cs.CL updates on arXiv.org) · arxiv.org
  • 2026-06-09: Decision-Aware Memory Cards: Counterfactual-Inspired Context Selection and Compression for Tool-Using LLM Agents (cs.AI updates on arXiv.org) · arxiv.org — arXiv Opus Qwen Codex GPT 5.5 Qwen-QLoRA Qwen3.6-Plus Gemini-3.1-Pro-High Qwen3.5-122B-A10B
  • 2026-06-09: AgentTrust: A Self-Improving Trust Layer for AI-Agent Actions (cs.AI updates on arXiv.org) · arxiv.org
  • 2026-06-09: Why Limit the Residual Stream to Layers and Not Tokens? Persistent Memory for Continuous Latent Reasoning (cs.AI updates on arXiv.org) · arxiv.org — arXiv GPT-2
  • 2026-06-09: MemoPilot: Training Memory Updates for LLM Agents with Reinforcement Learning (cs.CL updates on arXiv.org) · arxiv.org — arXiv DeepSeek-V3.2
  • 2026-06-09: Baichuan-M4: A Clinical-Grade Medical Agent System for Continuous Care (cs.AI updates on arXiv.org) · arxiv.org — Baichuan Intelligence Baichuan-M4
  • 2026-06-09: Syll: Open-Source Personal Automation with Cross-Surface Execution (cs.AI updates on arXiv.org) · arxiv.org — Adobe
  • 2026-06-09: MemToolAgent: Enhancing LLM Agent Tool Use Through Memory Management (cs.AI updates on arXiv.org) · arxiv.org — Danilo Neves Ribeiro
  • 2026-06-09: Rosetta Memory: Adaptive Memory for Cross-LLM Agents (cs.LG updates on arXiv.org) · arxiv.org
  • 2026-06-09: Hermes Codex Plugin: Local Memory for Coding Agents via SQLite (Искусственный интеллект – AI, ANN и иные формы искусственного разума) · habr.com
  • 2026-06-09: Navigating On-Premises LLM Deployment Challenges (Все статьи подряд / Искусственный интеллект / Хабр) · habr.com
  • 2026-06-09: Transitioning from SQL-prompts to multi-agent systems for team operations (Все статьи подряд / Искусственный интеллект / Хабр) · habr.com — OneCell AI Talent Hub Darya Voronkina

FAQ

What is Agent Memory?

Agent Memory is the set of mechanisms that lets AI agents persist useful context across turns, sessions, tools, and tasks. GROUNDING tracks memory stores, compression, retrieval boundaries, and failure modes for long-running agent workflows.

Which topic does Agent Memory belong to?

On the GROUNDING radar, Agent Memory is grouped under the Agents topic.

Related concepts tracked by the radar include Agents, Context Engineering, RAG.

Graph View

Backlinks