The Open-World Lottery Ticket Hypothesis for OOD Intent Classification
This work provides a method to enhance OOD detection in intent classification, which is incremental but beneficial for post hoc methods in natural language processing applications.
The paper tackles the problem of Out-of-Domain (OOD) intent classification by addressing model overconfidence, showing that pruning overparameterized models to find calibrated subnetworks improves confidence differentiation between in- and out-of-domain intents, with consistent improvements demonstrated on four real-world datasets.
Most existing methods of Out-of-Domain (OOD) intent classification rely on extensive auxiliary OOD corpora or specific training paradigms. However, they are underdeveloped in the underlying principle that the models should have differentiated confidence in In- and Out-of-domain intent. In this work, we shed light on the fundamental cause of model overconfidence on OOD and demonstrate that calibrated subnetworks can be uncovered by pruning the overparameterized model. Calibrated confidence provided by the subnetwork can better distinguish In- and Out-of-domain, which can be a benefit for almost all post hoc methods. In addition to bringing fundamental insights, we also extend the Lottery Ticket Hypothesis to open-world scenarios. We conduct extensive experiments on four real-world datasets to demonstrate our approach can establish consistent improvements compared with a suite of competitive baselines.