A Multi-cluster Boundary Learning Method for Out-of-Scope Intent Detection via MiniLM Embedding

基于 MiniLM 嵌入的多聚类边界学习方法用于范围外意图检测

Abstract: Intent detection is a critical task that bridges human intents and system actions in human-machine interaction systems. However, there still exist challenges for detecting out-of-scope (OOS) intents. (i) The traditional methods view the OOS intent detection as a multi-class classification, then the detection accuracy decreases as the class number of the known intents increases; (ii) LLM-embedding methods require large parameters, that makes them difficult to train and practically deploy.

摘要： 意图检测是人机交互系统中连接人类意图与系统行为的关键任务。然而，检测范围外（OOS）意图仍面临挑战：（i）传统方法将 OOS 意图检测视为多分类问题，随着已知意图类别的增加，检测准确率会下降；（ii）基于大语言模型（LLM）嵌入的方法参数量巨大，导致其难以训练和实际部署。

Thus, this work proposes a multi-cluster boundary learning method to detect OOS intents via MiniLM embedding (i.e., all-MiniLM-L6-v2) in an one-class classification workflow. The method learns the boundaries of multi-cluster embeddings generated by MiniLM from the training utterances, and then rejects the out-of-domain utterances as OOS intents.

因此，本研究提出了一种基于 MiniLM 嵌入（即 all-MiniLM-L6-v2）的多聚类边界学习方法，在单分类工作流中检测 OOS 意图。该方法通过训练话语学习由 MiniLM 生成的多聚类嵌入边界，进而将域外话语识别为 OOS 意图并予以拒绝。

Experiments are conducted on public CLINC150, StackOverflow and Banking77 datasets. The results show that the method achieves the state-of-the-art OOS intent detection performance compared the other baselines. Ablation studies are also conducted and the results show that the used MiniLM can better adapt to the workflow and utterance embedding requirements. The code is available at supplementary materials.

实验在公开的 CLINC150、StackOverflow 和 Banking77 数据集上进行。结果表明，与其他基线方法相比，该方法实现了最先进的 OOS 意图检测性能。消融实验结果进一步证明，所使用的 MiniLM 模型能更好地适应工作流及话语嵌入的需求。相关代码已在补充材料中提供。