arXiv cs.CL· Devarsh Gandhi, Sarthak Pawar, Hemprasad Badgujar, Vaibhav Rathore, Siddharth Sureka, Neha Bhargava·· 11 小时前AI 评分47
Scalaris：面向企业对话式 AI 助手的可扩展记忆架构
Scalaris: A Scalable Enterprise Memory Architecture for Conversational AI Assistants
AI 导读
Scalaris 是一种面向企业对话式 AI 助手的可扩展记忆架构，通过统一长短期记忆设计，在 Qwen3-32B 上取得 LongMemEval-S 92.2%、LoCoMo 90.77% 的 SOTA 成绩。其检索上下文 token 数比 Hindsight 少约 10 倍，接入其闭源 AI 助手后整体回复质量提升 36%。
正文 · 原文
Abstract:Enterprise conversational systems require memory to personalize interactions, retain information across sessions, and maintain conversational coherence. At enterprise scale, however, memory must not only retrieve relevant information accurately but also provide it efficiently to the answer-generation model. Existing memory systems face a trade-off between retrieval accuracy and context efficiency. Retrieving more context can improve recall but increases inference cost and latency. We present Scalaris, a scalable enterprise memory architecture that addresses this trade-off through a unified long-term and short-term memory design. We propose that user interactions and assistant responses possess distinct characteristics and roles in a conversation: user interactions capture evolving knowledge about the user and provide value across sessions, whereas assistant responses must be stored efficiently so that only information relevant to the current query is retrieved. We therefore treat them separately and support each with specialized representations and ingestion/retrieval pipelines. Scalaris reaches 92.2 % accuracy on LongMemEval-S and 90.77 % on LoCoMo using Qwen3-32B, state-of-the-art results on both benchmarks, while supplying the answer-generation model with approximately 10 times fewer retrieved context tokens than Hindsight, a leading memory system for conversational AI. When integrated with our proprietary closed-source AI assistants, Scalaris improved overall response quality by 36 %. Together, these results establish Scalaris as a cost-effective framework for building scalable, personalized conversational AI systems.
| Comments: | Paper Under Review |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.02612 [cs.CL] |
| (or arXiv:2611.02612v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.02612 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Vaibhav Rathore [view email]
[v1]
Mon, 2 Nov 2026 06:37:11 UTC (154 KB)
来源：arXiv cs.CL · arxiv.org