LycheeMemory V2：基於語義片段級鞏固的LLM智能體高效長期記憶

摘要

長程大語言模型代理必須保留過去互動中的資訊，以支援未來的任務。現有的記憶系統通常依賴急切式整合，在每次互動後呼叫大語言模型來提取、摘要或更新記憶。這種設計使得隨著對話增長，記憶建構的成本日益增加。粗略的摘要雖可降低建構成本，卻有丟失細粒度情境證據的風險；而擴大檢索上下文或採用多跳大語言模型推理，則會將開銷轉移至查詢階段。我們提出 LycheeMemory V2，一個以語意片段層級整合取代回合層級整合的高效長期記憶框架。LycheeMemory 不對每次互動進行整合，而是將多次交流批次化為片段，並將每個已定稿的片段編碼為與上下文無關的型別化記憶紀錄。片段層級批次處理可降低大語言模型的編碼頻率，而與固定視窗批次處理相比，語意邊界偵測有助於保留連貫的事件層級與時間證據。所產生的紀錄以輕量結構化索引組織，用於查詢規劃式證據檢索。使用 GPT-4.1-Mini 的實驗顯示，LycheeMemory 達到了最先進的效能，在 LoCoMo 上達到 89.22%，在 LongMemEval-S 上達到 92.20%。與 A-Mem 相比，它在 LoCoMo 上減少了 86.0% 的建構 token，在 LongMemEval-S 上減少了 75.9%，且未增加查詢階段的 token 使用量。更廣泛而言，我們的結果表明，長期代理記憶的準確度與成本之間的取捨，不僅取決於保留了哪些資訊，還取決於資訊整合的粒度。

English

Long-horizon LLM agents must preserve information from past interactions to support future tasks. Existing memory systems typically rely on eager consolidation, invoking LLMs after each interaction to extract, summarize, or update memories. This design makes memory construction increasingly costly as conversations grow. Coarse summarization can reduce construction cost but risks discarding fine-grained contextual evidence, whereas larger retrieval contexts or multi-hop LLM reasoning shift the overhead to query time. We present LycheeMemory V2, an efficient long-term memory framework that replaces turn-level consolidation with semantic segment-level consolidation. Instead of consolidating every interaction, LycheeMemory batches multiple exchanges into segments and encodes each finalized segment into context-independent typed memory records. Segment-level batching lowers LLM encoding frequency, while semantic boundary detection helps preserve coherent event-level and temporal evidence compared with fixed-window batching. The resulting records are organized with lightweight structured indexes for query-planned evidence retrieval. Experiments using GPT-4.1-Mini show that LycheeMemory achieves state-of-the-art performance, reaching 89.22% on LoCoMo and 92.20% on LongMemEval-S. Compared with A-Mem, it reduces construction tokens by 86.0% on LoCoMo and 75.9% on LongMemEval-S without increasing query-time token usage. More broadly, our results suggest that the accuracy--cost trade-off of long-term agent memory depends not only on what information is retained, but also on the granularity at which it is consolidated.