Data Augmentation: a Combined Inductive-Deductive Approach featuring Answer Set Programming
This addresses data scarcity issues in biomedical/healthcare applications, offering a method to augment datasets, though it appears incremental as it combines existing techniques like logic programming and deep learning.
The paper tackles the problem of limited and imbalanced datasets in domains like biomedical imaging by proposing a hybrid inductive-deductive approach that uses logic programs to generate structured labels and deep learning to create photo-realistic images, resulting in a framework for synthesizing labeled images that comply with domain constraints.
Although the availability of a large amount of data is usually given for granted, there are relevant scenarios where this is not the case; for instance, in the biomedical/healthcare domain, some applications require to build huge datasets of proper images, but the acquisition of such images is often hard for different reasons (e.g., accessibility, costs, pathology-related variability), thus causing limited and usually imbalanced datasets. Hence, the need for synthesizing photo-realistic images via advanced Data Augmentation techniques is crucial. In this paper we propose a hybrid inductive-deductive approach to the problem; in particular, starting from a limited set of real labeled images, the proposed framework makes use of logic programs for declaratively specifying the structure of new images, that is guaranteed to comply with both a set of constraints coming from the domain knowledge and some specific desiderata. The resulting labeled images undergo a dedicated process based on Deep Learning in charge of creating photo-realistic images that comply with the generated label.