Efficient Intent Detection with Dual Sentence Encoders
This work addresses the need for efficient and accessible intent detection in conversational AI, particularly for new domains with limited data, though it is incremental as it builds on existing encoder methods.
The paper tackled the problem of building resource-efficient intent detection models for conversational systems in low-data regimes by introducing methods using pretrained dual sentence encoders, showing they outperform BERT-based approaches on three datasets, especially with only 10 or 30 examples per intent, and can be trained in minutes on a CPU.
Building conversational systems in new domains and with added functionality requires resource-efficient models that work under low-data regimes (i.e., in few-shot setups). Motivated by these requirements, we introduce intent detection methods backed by pretrained dual sentence encoders such as USE and ConveRT. We demonstrate the usefulness and wide applicability of the proposed intent detectors, showing that: 1) they outperform intent detectors based on fine-tuning the full BERT-Large model or using BERT as a fixed black-box encoder on three diverse intent detection data sets; 2) the gains are especially pronounced in few-shot setups (i.e., with only 10 or 30 annotated examples per intent); 3) our intent detectors can be trained in a matter of minutes on a single CPU; and 4) they are stable across different hyperparameter settings. In hope of facilitating and democratizing research focused on intention detection, we release our code, as well as a new challenging single-domain intent detection dataset comprising 13,083 annotated examples over 77 intents.