Weakly Supervised Concept Map Generation through Task-Guided Graph Translation
This work addresses the need for interpretable knowledge summarization in NLP, offering a weakly supervised approach that is more efficient than deep generative models, though it is incremental in improving concept map generation techniques.
The paper tackles the problem of generating task-oriented concept maps from free texts by proposing GT-D2G, a framework that uses weak supervision from downstream task labels to translate initial graphs into concise structures, achieving better performance in document classification compared to other methods and demonstrating efficiency in label use.
Recent years have witnessed the rapid development of concept map generation techniques due to their advantages in providing well-structured summarization of knowledge from free texts. Traditional unsupervised methods do not generate task-oriented concept maps, whereas deep generative models require large amounts of training data. In this work, we present GT-D2G (Graph Translation-based Document To Graph), an automatic concept map generation framework that leverages generalized NLP pipelines to derive semantic-rich initial graphs, and translates them into more concise structures under the weak supervision of downstream task labels. The concept maps generated by GT-D2G can provide interpretable summarization of structured knowledge for the input texts, which are demonstrated through human evaluation and case studies on three real-world corpora. Further experiments on the downstream task of document classification show that GT-D2G beats other concept map generation methods. Moreover, we specifically validate the labeling efficiency of GT-D2G in the label-efficient learning setting and the flexibility of generated graph sizes in controlled hyper-parameter studies.