Sumedha Singla
YOU?
Author Swipe
View article: Subject2Vec: Generative-Discriminative Approach from a Set of Image Patches to a Vector
Subject2Vec: Generative-Discriminative Approach from a Set of Image Patches to a Vector Open
We propose an attention-based method that aggregates local image features to a subject-level representation for predicting disease severity. In contrast to classical deep learning that requires a fixed dimensional input, our method operate…
View article: Augmentation by Counterfactual Explanation -Fixing an Overconfident Classifier
Augmentation by Counterfactual Explanation -Fixing an Overconfident Classifier Open
A highly accurate but overconfident model is ill-suited for deployment in critical applications such as healthcare and autonomous driving. The classification outcome should reflect a high uncertainty on ambiguous in-distribution samples th…
View article: Augmentation by Counterfactual Explanation -- Fixing an Overconfident Classifier
Augmentation by Counterfactual Explanation -- Fixing an Overconfident Classifier Open
A highly accurate but overconfident model is ill-suited for deployment in critical applications such as healthcare and autonomous driving. The classification outcome should reflect a high uncertainty on ambiguous in-distribution samples th…
View article: Deep Learning for Medical Imaging From Diagnosis Prediction to its Counterfactual Explanation
Deep Learning for Medical Imaging From Diagnosis Prediction to its Counterfactual Explanation Open
Deep neural networks (DNN) have achieved unprecedented performance in computer-vision tasks almost ubiquitously in business, technology, and science. While substantial efforts are made to engineer highly accurate architectures and provide …
View article: Explaining the Black-box Smoothly- A Counterfactual Approach
Explaining the Black-box Smoothly- A Counterfactual Approach Open
We propose a BlackBox Counterfactual Explainer that is explicitly developed for medical imaging applications. Classical approaches (e.g., saliency maps) assessing feature importance do not explain how and why variations in a particular ana…
View article: Explaining the Black-box Smoothly- A Counterfactual Approach
Explaining the Black-box Smoothly- A Counterfactual Approach Open
We propose a BlackBox Counterfactual Explainer that is explicitly developed for medical imaging applications. Classical approaches (e.g., saliency maps) assessing feature importance do not explain how and why variations in a particular ana…
View article: Explaining the Black-box Smoothly- A Counterfactual Approach
Explaining the Black-box Smoothly- A Counterfactual Approach Open
We propose a BlackBox Counterfactual Explainer, designed to explain image classification models for medical applications. Classical approaches (e.g., saliency maps) that assess feature importance do not explain "how" imaging features in im…
View article: Incorporating External Information in Tissue Subtyping: A Topic Modeling Approach.
Incorporating External Information in Tissue Subtyping: A Topic Modeling Approach. Open
Probabilistic topic models, have been widely deployed for various applications such as learning disease or tissue subtypes. Yet, learning the parameters of such models is usually an ill-posed problem and may result in losing valuable infor…
View article: Improving clinical disease subtyping and future events prediction through a chest CT‐based deep learning approach
Improving clinical disease subtyping and future events prediction through a chest CT‐based deep learning approach Open
Purpose To develop and evaluate a deep learning (DL) approach to extract rich information from high‐resolution computed tomography (HRCT) of patients with chronic obstructive pulmonary disease (COPD). Methods We develop a DL‐based model to…
View article: Explanation by Progressive Exaggeration
Explanation by Progressive Exaggeration Open
As machine learning methods see greater adoption and implementation in high stakes applications such as medical image diagnosis, the need for model interpretability and explanation has become more critical. Classical approaches that assess…
View article: Subject2Vec: Generative-Discriminative Approach from a Set of Image\n Patches to a Vector
Subject2Vec: Generative-Discriminative Approach from a Set of Image\n Patches to a Vector Open
We propose an attention-based method that aggregates local image features to\na subject-level representation for predicting disease severity. In contrast to\nclassical deep learning that requires a fixed dimensional input, our method\noper…