发表机构
University of Exeter; Nanjing University; Shenzhen University; University of Oxford（埃克塞特大学; 南京大学; 深圳大学; 牛津大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
PersMem将个性整合到LLM代理的记忆处理流程中，通过四步双路径机制实现个性依赖的记忆处理，在依恋分类和Big Five对话中显著优于基线，验证了代理行为与预定义个性的一致性。
AI 中文摘要
角色扮演代理的画像通常依赖于系统提示中预定义的个性，而其记忆处理流程（包括存储记忆的优先级排序及后续检索）却独立于该个性。这种分离导致代理的记忆处理与预定义个性不一致，并且难以验证代理行为是否遵循该个性。在本文中，我们提出了个性整合记忆（Personality-Integrated Memory，PersMem），该方法将个性整合到代理的记忆处理流程中，使其始终依赖于个性。PersMem通过四个步骤处理记忆，其中个性被映射为操作特定参数，控制：（i）情感评估，标注用户输入的情绪状态；（ii）将先前存储的记忆与当前输入一起保留；（iii）被动情感驱动的记忆检索，探索与用户输入在语义和个性引导情绪上相似的记忆；以及（iv）主动目标驱动的记忆检索，对被动检索到的记忆进行细化和选择以生成回复。因此，可以通过检查人机交互过程中的记忆处理痕迹来检验与预定义个性的一致性。我们在依恋和Big Five设置中评估这些个性相关差异。PersMem在四类依恋分类中超过随机基线23.1个百分点。在Big Five对话比较中，PersMem达到67.5%的准确率，比使用均匀采样记忆的基线高出6.7个百分点。在CoSER上，PersMem的平均得分为66.13，其中角色保真度得分为69.33，故事情节质量得分为84.33。这些结果表明，PersMem产生了可区分的个性相关记忆处理模式。
英文摘要
The profile of a role-playing agent usually depends on the pre-defined personality in a system prompt, whereas its memory processing pipeline, including prioritisation of stored memories and subsequent retrieval, remains independent of this personality. This separation causes the agent's memory processing to be inconsistent with the pre-defined personality, and makes it difficult to validate whether agent behaviours follow this personality. In this paper, we propose Personality-Integrated Memory (PersMem), which integrates personality into the agent's memory processing pipeline, making it consistently personality-dependent. PersMem processes memory using four steps, where the personality is mapped to operation-specific parameters controlling: (i) affective appraisal annotating emotion states of the user input; (ii) retention of previously stored memories along with the current input; (iii) passive affect-driven memory retrieval exploring memories similar to user input in semantics and personality-guided emotions; and (iv) active goal-driven memory retrieval that refines and selects passively retrieved memories for the reply. Consequently, consistency with the pre-defined personality can be examined by inspecting memory-processing traces during human-agent interactions. We evaluate these personality-dependent differences in attachment and Big Five settings. PersMem exceeds the chance baseline for four-way attachment classification by 23.1 percentage points. In Big Five dialogue comparisons, PersMem achieves 67.5% accuracy, 6.7 percentage points above a baseline using uniformly sampled memories. On CoSER, PersMem achieves an average score of 66.13, with scores of 69.33 for Character Fidelity and 84.33 for Storyline Quality. Together, these results show that PersMem produces distinguishable personality-related memory-processing patterns.
Comments39 pages, 2 figures
