Seed-Guided Topic Discovery with Out-of-Vocabulary Seeds
This work addresses a domain-specific issue in natural language processing by enhancing topic modeling for users with specific interests, though it is incremental as it builds on existing seed-guided approaches.
The paper tackles the problem of seed-guided topic discovery by allowing out-of-vocabulary seeds and leveraging pre-trained language models, resulting in improved topic coherence, accuracy, and diversity on three real datasets.
Discovering latent topics from text corpora has been studied for decades. Many existing topic models adopt a fully unsupervised setting, and their discovered topics may not cater to users' particular interests due to their inability of leveraging user guidance. Although there exist seed-guided topic discovery approaches that leverage user-provided seeds to discover topic-representative terms, they are less concerned with two factors: (1) the existence of out-of-vocabulary seeds and (2) the power of pre-trained language models (PLMs). In this paper, we generalize the task of seed-guided topic discovery to allow out-of-vocabulary seeds. We propose a novel framework, named SeeTopic, wherein the general knowledge of PLMs and the local semantics learned from the input corpus can mutually benefit each other. Experiments on three real datasets from different domains demonstrate the effectiveness of SeeTopic in terms of topic coherence, accuracy, and diversity.