ML model design either starts with an interpretable model or a Blackbox and explains it post hoc. Blackbox models are flexible but difficult to explain, while interpretable models are inherently explainable. Yet, interpretable models require extensive ML knowledge and tend to be less flexible, potentially underperforming than their Blackbox equivalents. This paper aims to blur the distinction between a post hoc explanation of a Blackbox and constructing interpretable models. Beginning with a Blackbox, we iteratively carve out a mixture of interpretable models and a residual network. The interpretable models identify a subset of samples and explain them using First Order Logic (FOL), providing basic reasoning on concepts from the Blackbox. We route the remaining samples through a flexible residual. We repeat the method on the residual network until all the interpretable models explain the desired proportion of data. Our extensive experiments show that our route, interpret, and repeat approach (1) identifies a richer diverse set of instance-specific concepts with high concept completeness via interpretable models by specializing in various subsets of data without compromising in performance, (2) identifies the relatively “harder” samples to explain via residuals, (3) outperforms the interpretable by-design models by significant margins during test-time interventions, (4) can be used to fix the shortcut learned by the original Blackbox.

Generative Adversarial Networks (GAN) have many potential medical imaging applications, including data augmentation, domain adaptation, and model explanation. Due to the limited memory of Graphical Processing Units (GPUs), most current 3D GAN models are trained on low-resolution medical images, these models either cannot scale to high-resolution or are prone to patchy artifacts. In this work, we propose a novel end-to-end GAN architecture that can generate high-resolution 3D images. We achieve this goal by using different configurations between training and inference. During training, we adopt a hierarchical structure that simultaneously generates a low-resolution version of the image and a randomly selected sub-volume of the high-resolution image. The hierarchical design has two advantages: First, the memory demand for training on high-resolution images is amortized among sub-volumes. Furthermore, anchoring the high-resolution sub-volumes to a single low-resolution image ensures anatomical consistency between sub-volumes. During inference, our model can directly generate full high-resolution images. We also incorporate an encoder with a similar hierarchical structure into the model to extract features from the images. Experiments on 3D thorax CT and brain MRI demonstrate that our approach outperforms state of the art in image generation. We also demonstrate clinical applications of the proposed model in data augmentation and clinical-relevant feature extraction.

Supervised learning method requires a large volume of annotated datasets. Collecting such datasets is time-consuming and expensive. Until now, very few annotated COVID-19 imaging datasets are available. Although self-supervised learning enables us to bootstrap the training by exploiting unlabeled data, the generic self-supervised methods for natural images do not sufficiently incorporate the context. For medical images, a desirable method should be sensitive enough to detect deviation from normal-appearing tissue of each anatomical region; here, anatomy is the context. We introduce a novel approach with two levels of self-supervised representation learning objectives: one on the regional anatomical level and another on the patient-level. We use graph neural networks to incorporate the relationship between different anatomical regions. The structure of the graph is informed by anatomical correspondences between each patient and an anatomical atlas. In addition, the graph representation has the advantage of handling any arbitrarily sized image in full resolution. Experiments on large-scale Computer Tomography (CT) datasets of lung images show that our approach compares favorably to baseline methods that do not account for the context. We use the learnt embedding to quantify the clinical progression of COVID-19 and show that our method generalizes well to COVID-19 patients from different hospitals. Qualitative results suggest that our model can identify clinically relevant regions in the images.

Extracting hidden phenotypes is essential in medical data analysis because it facilitates disease subtyping, diagnosis, and understanding of disease etiology. Since the hidden phenotype is usually a low-dimensional representation that comprehensively describes the disease, we require a dimensionality reduction method that captures as much disease-relevant information as possible. However, most unsupervised or self-supervised methods cannot achieve the goal because they learn a holistic representation containing both disease-relevant and disease-irrelevant information. Supervised methods can capture information that is predictive to the target clinical variable only, but the learned representation is usually not generalizable for the various aspects of the disease. Hence, we develop a dimensionality-reduction approach to extract Disease Relevant Features (DRFs) based on information theory. We propose to use clinical variables that weakly define the disease as so-called anchors. We derive a formulation that makes the DRF predictive of the anchors while forcing the remaining representation to be irrelevant to the anchors via adversarial regularization. We apply our method to a large-scale study of Chronic Obstructive Pulmonary Disease (COPD). Our experiment shows: (1) Learned DRFs are as predictive as the original representation in predicting the anchors, although it is in a significantly lower dimension. (2) Compared to supervised representation, the learned DRFs are more predictive to other relevant disease metrics that are not used during the training. (3) The learned DRFs are related to non-imaging biological measurements such as gene expressions, suggesting the DRFs include information related to the underlying biology of the disease.

The generalization of representations learned via contrastive learning depends crucially on what features of the data are extracted. However, we observe that the contrastive loss does not always sufficiently guide which features are extracted, a behavior that can negatively impact the performance on downstream tasks via “shortcuts", i.e., by inadvertently suppressing important predictive features. We find that feature extraction is influenced by the difficulty of the so-called instance discrimination task (i.e., the task of discriminating pairs of similar points from pairs of dissimilar ones). Although harder pairs improve the representation of some features, the improvement comes at the cost of suppressing previously well represented features. In response, we propose implicit feature modification (IFM), a method for altering positive and negative samples in order to guide contrastive models towards capturing a wider variety of predictive features. Empirically, we observe that IFM reduces feature suppression, and as a result improves performance on vision and medical imaging tasks.