FastClass: A Time-Efficient Approach to Weakly-Supervised Text Classification

FastClass: A Time-Efficient Approach to Weakly-Supervised Text Classification Tingyu Xia1,3, Yue Wang2, Yuan Tian1,3,∗, Yi Chang1,3,4,∗ 1School of Artificial Intelligence, Jilin University 2School of Information and Library Science, University of North Carolina at Chapel Hill 3Key Laboratory of Symbolic Computation and Knowledge Engineering, Jilin University 4International Center of Future Science, Jilin University xiaty21@mails.jlu.edu.cn, wangyue@unc.edu yuantian@jlu.edu.cn, yichang@jlu.edu.cn Abstract Weakly-supervised text classification aims to train a classifier using only class descriptions and unlabeled data. Recent research shows that keyword-driven methods can achieve state-of- the-art performance on various tasks. How- ever, these methods not only rely on carefully- crafted class descriptions to obtain class- specific keywords but also require substantial amount of unlabeled data and takes a long time to train. This paper proposes FastClass, an efficient weakly-supervised classification ap- proach. It uses dense text representation to retrieve class-relevant documents from external unlabeled corpus and selects an optimal sub- set to train a classifier. Compared to keyword- driven methods, our approach is less reliant on initial class descriptions as it no longer needs to expand each class description into a set of class-specific keywords. Experiments on a wide range of classification tasks show that the proposed approach frequently outper- forms keyword-driven models in terms of clas- sification accuracy and often enjoys orders-of- magnitude faster training speed. 1 Introduction Text classification is one of the most used tech- niques in mining large-scale unstructured text. When sufficient labeled data are available, super- vised classification techniques can achieve excel- lent performance. However, manually labeling example documents can be time-consuming and labor-intensive, a major burden when applying su- pervised text classification techniques in practice. Recently, weakly-supervised text classifica- tion (Meng et al., 2018, 2019, 2020; Mekala and Shang, 2020; Shen et al., 2021; Wang et al., 2021; Zhang et al., 2021) has been proposed to save la- beling efforts. It refers to the ability for a machine learning model to start classifying documents by using only class descriptions and unlabeled data. ∗Corresponding authors. Since weakly-supervised text classification meth- ods do not rely on any labeled data, it can greatly reduce the workload of data labeling. Therefore, these methods are desirable in real-world text min- ing applications. The current mainstream weakly-supervised text classification methods are keyword-driven (Meng et al., 2018, 2019, 2020; Mekala and Shang, 2020; Shen et al., 2021; Wang et al., 2021; Zhang et al., 2021), where the users need to carefully choose initial class-relevant keywords for each class. Such keywords are expanded into a richer set of key- words, which are then used to assign pseudo- labels on unlabeled data. The critical step of these keyword-driven methods is the choice of initial class-relevant keywords and the subsequent key- word expansion strategy, which determines the quality of pseudo-labeled documents. However, if the user’s initial choice of class de- scriptions is ambiguous or brief, then the results of keyword expansion will be negatively impacted. Ideally, the user can provide initial class descrip- tions that are beneficial to keywords expansion. However, in reality, the quality of class descriptions …

Content truncated to 500 words