Interactive Supervision for New Intent Discovery

• Published in: IEEE signal processing letters Vol. 31; pp. 1 - 5
• Main Authors: Hu, Zhanxuan, Xu, Yan, He, Lang, Nie, Feiping
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.01.2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Banking; Clustering; Clusters; embedding; Intent discovery; Learning; Measurement; Reliability; Representation learning; Representations; Semantics; Supervision; Training; unsupervised learning; Vectors
• ISSN: 1070-9908; 1558-2361
• DOI: 10.1109/LSP.2024.3416882

New Intent Discovery (NID) is the task of categorizing new and unknown intents into distinct clusters. Recent advances in this issue can be roughly grouped into parametric clustering and representation learning. Although both of them have demonstrated exceptional performance, the efficacy of combining these two techniques remains unexplored. To this end, we propose a new framework named INS-NID ( IN teractive S upervision for N ew I ntent D iscovery). This framework is designed to build a bridge between parametric clustering and representation learning. In practice, INS-NID comprises a parametric clustering branch and a representation learning branch, which work collaboratively and provide interactive supervision to boost each other. Specifically, the representation learning branch provides reliable distribution estimation during training, which is used to regularize the cluster assignments of the parametric clustering. On the other hand, the cluster assignments predicted by parametric clustering provide additional supervision information for the representation learning branch. The effectiveness of INS is highlighted with superior performance over several state-of-the-art methods across various benchmarks. For example, INS achieves 90% clustering accuracy on the Banking dataset, surpassing the best competitor by 5.96%. Our code is available at https://github.com/Tarrius/INS/ .