RoNID: New Intent Discovery with Generated-Reliable Labels and Cluster-friendly Representations

arXiv:2404.08977v1 Announce Type: cross
Abstract: New Intent Discovery (NID) strives to identify known and reasonably deduce novel intent groups in the open-world scenario. But current methods face issues with inaccurate pseudo-labels and poor representation learning, creating a negative feedback loop that degrades overall model performance, including accuracy and the adjusted rand index. To address the aforementioned challenges, we propose a Robust New Intent Discovery (RoNID) framework optimized by an EM-style method, which focuses on constructing reliable pseudo-labels and obtaining cluster-friendly …

abstract accuracy arxiv cluster cs.cl cs.lg current discovery face feedback generated identify labels loop negative novel open-world performance representation representation learning type world