Search Results for author: Divya Shanmugam

Found 10 papers, 2 papers with code

Improved Text Classification via Test-Time Augmentation

no code implementations27 Jun 2022 Helen Lu, Divya Shanmugam, Harini Suresh, John Guttag

Test-time augmentation -- the aggregation of predictions across transformed examples of test inputs -- is an established technique to improve the performance of image classification models.

Classification Image Classification +2

Data Augmentation for Electrocardiograms

1 code implementation9 Apr 2022 Aniruddh Raghu, Divya Shanmugam, Eugene Pomerantsev, John Guttag, Collin M. Stultz

In experiments, considering three datasets and eight predictive tasks, we find that TaskAug is competitive with or improves on prior work, and the learned policies shed light on what transformations are most effective for different tasks.

Data Augmentation

Quantifying Inequality in Underreported Medical Conditions

1 code implementation8 Oct 2021 Divya Shanmugam, Emma Pierson

Estimating the prevalence of a medical condition, or the proportion of the population in which it occurs, is a fundamental problem in healthcare and public health.

Better Aggregation in Test-Time Augmentation

no code implementations ICCV 2021 Divya Shanmugam, Davis Blalock, Guha Balakrishnan, John Guttag

In this paper, we present 1) experimental analyses that shed light on cases in which the simple average is suboptimal and 2) a method to address these shortcomings.

Image Classification

Unsupervised Domain Adaptation in the Absence of Source Data

no code implementations20 Jul 2020 Roshni Sahoo, Divya Shanmugam, John Guttag

Current unsupervised domain adaptation methods can address many types of distribution shift, but they assume data from the source domain is freely available.

Unsupervised Domain Adaptation

Multiple Instance Learning for ECG Risk Stratification

no code implementations2 Dec 2018 Divya Shanmugam, Davis Blalock, John Guttag

We focus on estimating a patient's risk of cardiovascular death after an acute coronary syndrome based on a patient's raw electrocardiogram (ECG) signal.

Ecg Risk Stratification Multiple Instance Learning

An Energy-Based Framework for Arbitrary Label Noise Correction

no code implementations27 Sep 2018 Jaspreet Sahota, Divya Shanmugam, Janahan Ramanan, Sepehr Eghbali, Marcus Brubaker

We propose an energy-based framework for correcting mislabelled training examples in the context of binary classification.

Cannot find the paper you are looking for? You can Submit a new open access paper.