Recall and personalization, enabled by memory, are pivotal to a lifelong conversational agent. 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. To solve this, we introduce SaliMory, a framework that learns to manage and utilize a cognitively structured memory spanning user facts, preferences, and working memory. By introducing novel stage wise process rewards and reward decomposed contrastive refinement, SaliMory provides isolated supervision for distinct memory operations (selective filtering, consolidation, and cue driven recall) end to end. Across four datasets, SALIMORY records the highest Good Personalization rate at 39.8% and the best QA accuracy at 72.9%, while consistently producing the least vague and most factual memories, keeping vagueness under 20% and factuality above 89%.

翻译：暂无翻译