Enriched Pre-trained Transformers for Joint Slot Filling and Intent Detection
This work addresses data scarceness and specialized vocabulary challenges in natural language understanding for applications like virtual assistants, representing an incremental improvement over existing BERT-based methods.
The paper tackled joint slot filling and intent detection in natural language understanding by proposing a novel architecture based on BERT and RoBERTa, incorporating an intent pooling attention mechanism and fusing intent distributions with word features, which outperformed current non-BERT state-of-the-art and some BERT-based baselines on standard datasets.
Detecting the user's intent and finding the corresponding slots among the utterance's words are important tasks in natural language understanding. Their interconnected nature makes their joint modeling a standard part of training such models. Moreover, data scarceness and specialized vocabularies pose additional challenges. Recently, the advances in pre-trained language models, namely contextualized models such as ELMo and BERT have revolutionized the field by tapping the potential of training very large models with just a few steps of fine-tuning on a task-specific dataset. Here, we leverage such models, namely BERT and RoBERTa, and we design a novel architecture on top of them. Moreover, we propose an intent pooling attention mechanism, and we reinforce the slot filling task by fusing intent distributions, word features, and token representations. The experimental results on standard datasets show that our model outperforms both the current non-BERT state of the art as well as some stronger BERT-based baselines.