Research

Research Interests:

• Applied Deep Learning and Data Science
• Machine Learning based Computer Vision: image classification, image quality assessment, image super resolution, image segmentation, generative models (GAN, VAE, etc.)
• Natural Language Processing: text classification, sentiment analysis, question answering systems, NER, text auto-moderation

Recent Projects

• User generated content moderation at Expedia Inc.
• User search query analysis from NLP perspective, including type classification (with CNN and LSTM), for name entity recognition, and topic modeling
• Learning domain specific embedding for words and entities, by context aware models.

Past Projects

• Video analysis for personalized advertisement using deep audio-visual convolutional network (joint work with Samsung Research America)
• Deep learning based question-answering systems (joint work with AT&T research lab).
• Medical image classification using deep unsupervised learning (joint work with NYU Medical School):
  • Adversarial Auto-encoders are used to learn feature representation from MR Images, and are used for classification of diseases.
• Social Media Video Summarization Using Deep Features (in collaboration with Verizon). Winner of Verizon Open Innovation Challenge.
• Image Segmentation Using Robust Regression and Sparse Decomposition, and its application for High Efficiency Video Coding Extensions (joint work with Huawei Labs)
• Scattering Convolutional Network for Image Classification and Biometrics Recognition (including face, iris, fingerprint, palmprint, etc.)
• Prediction of Performance of People on Neuropsychology Tests Using fMRI Features

Resume:

• For more details please refer to my resume here.