Inducing Meaningful Units from Character Sequences with Dynamic Capacity Slot Attention
This addresses the challenge of unsupervised unit discovery in text for natural language processing, but it appears incremental as it extends an existing architecture from images to sequences.
The paper tackles the problem of learning abstract meaningful units from character sequences without supervision, proposing a Dynamic Capacity Slot Attention model that discovers continuous representations of objects in sequences, and experiments show it discovers units similar to previous proposals in form, content, and abstraction level.
Characters do not convey meaning, but sequences of characters do. We propose an unsupervised distributional method to learn the abstract meaningful units in a sequence of characters. Rather than segmenting the sequence, our Dynamic Capacity Slot Attention model discovers continuous representations of the objects in the sequence, extending an architecture for object discovery in images. We train our model on different languages and evaluate the quality of the obtained representations with forward and reverse probing classifiers. These experiments show that our model succeeds in discovering units which are similar to those proposed previously in form, content and level of abstraction, and which show promise for capturing meaningful information at a higher level of abstraction.