Improved Out-of-Scope Intent Classification with Dual Encoding and Threshold-based Re-Classification

TL;DR

The paper introduces DETER, a novel end-to-end framework using dual encoders and threshold-based re-classification to improve out-of-scope intent detection without distribution assumptions, achieving state-of-the-art results.

Contribution

DETER employs dual text encoders and synthetic outlier generation to enhance out-of-scope intent detection in dialogue systems.

Findings

Outperforms previous benchmarks on CLINC-150, Stackoverflow, and Banking77 datasets.

Increases F1 score by up to 13% and 5% for known and unknown intents on CLINC-150.

Achieves 16% and 24% improvements in F1 scores for known and unknown intents on Banking77.

Abstract

Detecting out-of-scope user utterances is essential for task-oriented dialogues and intent classification. Current methodologies face difficulties with the unpredictable distribution of outliers and often rely on assumptions about data distributions. We present the Dual Encoder for Threshold-Based Re-Classification (DETER) to address these challenges. This end-to-end framework efficiently detects out-of-scope intents without requiring assumptions on data distributions or additional post-processing steps. The core of DETER utilizes dual text encoders, the Universal Sentence Encoder (USE) and the Transformer-based Denoising AutoEncoder (TSDAE), to generate user utterance embeddings, which are classified through a branched neural architecture. Further, DETER generates synthetic outliers using self-supervision and incorporates out-of-scope phrases from open-domain datasets. This approach ensures a comprehensive training set for out-of-scope detection. Additionally, a threshold-based re-classification mechanism refines the model's initial predictions. Evaluations on the CLINC-150, Stackoverflow, and Banking77 datasets demonstrate DETER's efficacy. Our model outperforms previous benchmarks, increasing up to 13% and 5% in F1 score for known and unknown intents on CLINC-150 and Stackoverflow, and 16% for known and 24% % for unknown intents on Banking77. The source code has been released at https://github.com/Hossam-Mohammed-tech/Intent_Classification_OOS.