让大模型在对话中零样本填充槽位，提升准确率与实时性
Zero-shot Slot Filling in the Age of LLMs for Dialogue Systems
- 用自动标注和知识蒸馏，把大模型能力迁移到小模型
- 在内部数据上比直接用大模型高26%的F1分数
- 适合需要低延迟高精度的客服对话系统
零样本槽位填充是自然语言理解中的经典任务。然而，现有方法多集中于单轮文本，忽视了对话数据特有的动态性：话题突变、打断和隐含指代等问题，使大语言模型（LLMs）的卓越能力难以直接应用。本文提出自动数据标注与黑盒知识蒸馏策略，将教师模型的知识迁移至小型模型，在内部数据集上相比原生大模型实现了26%的绝对F1提升。此外，我们设计了一种适用于呼叫中心产品场景的高效系统架构，相较现成抽取模型相对提升34% F1，支持对对话流进行近实时推理，兼具高精度与低延迟。
原文摘要 · Abstract (English)
Zero-shot slot filling is a well-established subtask of Natural Language Understanding (NLU). However, most existing methods primarily focus on single-turn text data, overlooking the unique complexities of conversational dialogue. Conversational data is highly dynamic, often involving abrupt topic shifts, interruptions, and implicit references that make it difficult to directly apply zero-shot slot filling techniques, even with the remarkable capabilities of large language models (LLMs). This paper addresses these challenges by proposing strategies for automatic data annotation with slot induction and black-box knowledge distillation (KD) from a teacher LLM to a smaller model, outperforming vanilla LLMs on internal datasets by 26% absolute increase in F1 score. Additionally, we introduce an efficient system architecture for call center product settings that surpasses off-the-shelf extractive models by 34% relative F1 score, enabling near real-time inference on dialogue streams with higher accuracy, while preserving low latency.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。