Tag
研究提出按语义类别条件化的记忆保留置信阈值，而非单一全局阈值，用于智能体长期记忆的写入边界控制，在Meta部署的冷启动记忆管线中将不受支持的保留率从6.2%降至4.0%。
The article asks AI users to share real examples of context forgetting in tools like ChatGPT, Claude, and Gemini, with the aim of researching the problem before building new solutions.
Analysts and JPMorgan Chase identify memory retention in humanoid robots as the next major breakthrough and a key bottleneck in their development.
This paper introduces CAMeR, a memory retention framework for LLM agents that combines keyword-gated hybrid activation with adaptive weight dynamics, and presents CAMeR-Bench, a benchmark for evaluating adaptive memory retention. Experiments show that hybrid symbolic-neural gating improves retention gaps and retrieval efficiency compared to embedding-only or time-driven baselines.
This paper formulates memory retention for long-horizon language agents as a constrained stochastic optimization problem, introducing OSL-MR, a framework that enforces observability-safe learning with a Mixed-Score heuristic. Experiments show consistent improvements over existing heuristic baselines under tight memory budgets.
The paper introduces Self-Consolidating Language Models (SCoL), a framework that uses meta-reinforcement learning to write current context into model weights for continual knowledge incorporation. It demonstrates improved acquisition and retention over baselines in both QA and long-context consolidation tasks.