Career Profile

I’m a PhD student under the aegis of Dr. Anna Choromanska at the Machine Learning Lab, Department of Electrical Engineering, New York University.

Professional Experience

Machine Learning Intern

Summer, 2018
Hearst, New York
  • Incorporated content based image retrieval into Hearst’s digital asset management system consisting of over 13 million images. We utilized a product quantizer based nearest neighbor search on image features extracted from deep models pre-trained on ImageNet.
  • Developed an efficient image visualization and tagging tool that utilizes t-distributed stochastic neighbor embedding (t-SNE).

Research experience

Estimating Sample Complexity in Deep Neural Networks
  • The most common and effective strategy in deep learning to improve performance and reduce generalization error is to collect more training data \textit{but} the fundamental question: “how much data do we need to achieve certain desirable performance?” still remains open. We aim to present a practical and predictive probabilistic method to model the relationship between training data sizes and misclassification rate.
VisualBackProp for learning using privileged information with CNNs
  • Proposed a novel training paradigm for CNNs based on learning using privileged information (LUPI) previously proposed for SVMs. Priviledge information (segmentation mask) was incorporated into training by additional loss functions to focus model’s attention on the desired region of the input image. Proposed method was verified through experiments on ImageNet and PASCAL VOC data sets and performance improvements of 2.4% and 2.7% over standard single-supervision model were obtained.
Towards Automated Melanoma Detection with Deep Learning:Data Purification and Augmentation
  • Proposed a data preparation strategy to tackle detrimental effects of biased data-sets on deep learning models. We utilized partial convolutions for removal of occlusions and synthetic data generation through coupled GAN to balance disease classes. We demonstrate its effectiveness on ISIC 2017 and 2018 grand challenge.
High Frequency Ultrasound Image Segmentation and Analysis
  • Trained an Active Shape Model to segment brain ventricles of a mouse embryo from its high frequency 3D Ultrasound image. The shape of the brain ventricle was described using a shape context descriptor while principle component analysis was used to generate the model.


Towards Automated Melanoma Detection with Deep Learning Data Purification and Augmentation.
Devansh Bisla, Anna Choromanska, Jennifer A. Stein, David Polsky, Russell Berman
CVPR - ISIC, 2019
VisualBackProp for learning using privileged information with CNNs
Devansh Bisla, Anna Choromanska
Arxiv, 2018

Skills & Proficiency