Pre-training Intent-aware Encoders For Zero- And Few-shot Intent Classification

Sung Mujeen, Gung James, Mansimov Elman, Pappas Nikolaos, Shu Raphael, Romeo Salvatore, Zhang Yi, Castelli Vittorio. Arxiv 2023

[Paper]    
Few Shot Pretraining Methods Reinforcement Learning Training Techniques

Intent classification (IC) plays an important role in task-oriented dialogue systems. However, IC models often generalize poorly when training without sufficient annotated examples for each user intent. We propose a novel pre-training method for text encoders that uses contrastive learning with intent psuedo-labels to produce embeddings that are well-suited for IC tasks, reducing the need for manual annotations. By applying this pre-training strategy, we also introduce Pre-trained Intent-aware Encoder (PIE), which is designed to align encodings of utterances with their intent names. Specifically, we first train a tagger to identify key phrases within utterances that are crucial for interpreting intents. We then use these extracted phrases to create examples for pre-training a text encoder in a contrastive manner. As a result, our PIE model achieves up to 5.4% and 4.0% higher accuracy than the previous state-of-the-art text encoder for the N-way zero- and one-shot settings on four IC datasets.

Similar Work