UCTopic: Unsupervised Contrastive Learning for Phrase Representations and Topic Mining
This addresses the need for context-aware phrase representations in topic mining without extensive annotations, though it appears incremental as it builds on contrastive learning with novel components.
The paper tackles the problem of learning high-quality phrase representations for topic mining by proposing UCTopic, an unsupervised contrastive learning framework that improves phrase representations and achieves a 38.2% average NMI gain over state-of-the-art models on entity clustering tasks.
High-quality phrase representations are essential to finding topics and related terms in documents (a.k.a. topic mining). Existing phrase representation learning methods either simply combine unigram representations in a context-free manner or rely on extensive annotations to learn context-aware knowledge. In this paper, we propose UCTopic, a novel unsupervised contrastive learning framework for context-aware phrase representations and topic mining. UCTopic is pretrained in a large scale to distinguish if the contexts of two phrase mentions have the same semantics. The key to pretraining is positive pair construction from our phrase-oriented assumptions. However, we find traditional in-batch negatives cause performance decay when finetuning on a dataset with small topic numbers. Hence, we propose cluster-assisted contrastive learning(CCL) which largely reduces noisy negatives by selecting negatives from clusters and further improves phrase representations for topics accordingly. UCTopic outperforms the state-of-the-art phrase representation model by 38.2% NMI in average on four entity cluster-ing tasks. Comprehensive evaluation on topic mining shows that UCTopic can extract coherent and diverse topical phrases.