SaliMory: Orchestrating Cognitive Memory for Conversational Agents

SaliMory：为对话智能体编排认知记忆

Abstract: Conversational agents that serve as lifelong companions must maintain persistent memory across all interactions. However, simply expanding context windows with raw retrieval degrades reasoning quality, while training memory agents via standard reinforcement learning creates a severe credit assignment bottleneck in a multi-stage pipeline.

摘要： 作为终身伴侣的对话智能体必须在所有交互过程中保持持久的记忆。然而，仅仅通过原始检索来扩展上下文窗口会降低推理质量，而通过标准强化学习训练记忆智能体则会在多阶段流水线中造成严重的信用分配瓶颈。

To solve this, we introduce SALIMORY, a framework that trains a single language model to manage a cognitively-structured memory-spanning user facts, preferences, and working memory. By introducing a hierarchical stage-wise process reward and reward-decomposed contrastive refinement, SALIMORY provides isolated supervision for distinct memory operations (selective filtering, consolidation, and cue-driven recall) end-to-end.

为了解决这一问题，我们引入了 SALIMORY，这是一个训练单一语言模型来管理认知结构化记忆的框架，涵盖了用户事实、偏好和工作记忆。通过引入分层阶段式过程奖励和奖励分解对比细化，SALIMORY 为不同的记忆操作（选择性过滤、巩固和线索驱动的回忆）提供了端到端的独立监督。

SALIMORY 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.

SALIMORY 将与记忆相关的故障减少了三分之一，端到端准确率比现有最先进技术提高了 10% 以上，并将“良好个性化”（Good Personalization）率提高了一倍以上。