Bio

I am a third year PhD student in the Computer Science Department at Rutgers University. I have been fortunate enough to work with my advisor, Prof. Vladimir Pavlovic. I obtained my M.S. degree from the Computer Science Department at Rutgers University and B.Tech from National Institute of Technology, Rourkela, India. My research interest is in applied side of machine learning and computer vision. In particular, I work on problems pertaining to transfer learning and disentanglement of latent features deep learning using deep learning.

Conference, Journal and Workshops

Unpacking Large Language Models with Conceptual Consistency Pritish Sahu*, Michael Cogswell, Yunye Gong, Ajay Divakaran arXiv preprint [pdf]

DAReN: A Collaborative Approach Towards Visual Reasoning And Disentangling Pritish Sahu*, Kalliopi Basioti, Vladimir Pavlovic In Proceedings of International Conference on Pattern Recognition (ICPR, 2022, Oral) [pdf]

SAViR-T: Spatially Attentive Visual Reasoning with Transformers Pritish Sahu*, Kalliopi Basioti, Vladimir Pavlovic In Proceedings of European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD, 2022, Oral) [pdf]

Challenges in Procedural Multimodal Machine Comprehension: A Novel Way To Benchmark Pritish Sahu*, Karan Sikka*, Ajay Divakaran In Proceedings of Winter Conference on Applications of Computer Vision (WACV, 2022) [pdf]

Comprehension Based Question Answering using Bloom’s Taxonomy Pritish Sahu*, Michael Cogswell, Sarah Rutherford-Quach, Ajay Divakaran Proceedings of the 6th Workshop on Representation Learning for NLP, Association for Computational Linguistics (RepL4NLP-2021, ACL, 2021) [pdf]

Towards Solving Multimodal Comprehension Pritish Sahu*, Karan Sikka*, Ajay Divakaran ArXiv preprint [pdf]

Zero-Shot Learning with Knowledge Enhanced Visual Semantic Embeddings Karan Sikka, Jihua Huang, Andrew Silberfarb, Prateeth Nayak, Luke Rohrer, Pritish Sahu, John Byrnes, Ajay Divakaran, Richard Rohwer ArXiv preprint [pdf]

Bayes-Factor-VAE: Hierarchical Bayesian Deep Auto-Encoder Models for Factor Disentanglement Minyoung Kim, Yuting Wang*, Pritish Sahu*, Vladimir Pavlovic In Proceedings of International Conference of Computer Vision (ICCV 2019, Oral) [pdf]

Task-Discriminative Domain Alignment for Unsupervised Domain Adaptation Behnam Gholami, Pritish Sahu, Minyoung Kim, Vladimir Pavlovic In Proceedings of International Conference of Computer Vision (MDALC Workshop at ICCV 2019, Oral) [pdf]

Unsupervised visual domain adaptation: A deep max-margin gaussian process approach Minyoung Kim, Pritish Sahu, Behnam Gholami, Vladimir Pavlovic In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2019, Oral) [pdf] [Video]

Relevance Factor VAE: Learning and Identifying Disentangled Factors Minyoung Kim, Yuting Wang*, Pritish Sahu*, Vladimir Pavlovic [pdf]

Unsupervised Multi-Target Domain Adaptation: An Information Theoretic Approach Behnam Gholami, Pritish Sahu, Ognjen Rudovic, Konstantinos Bousmalis, Vladimir Pavlovic [pdf]

Filling in the blanks: reconstructing microscopic crowd motion from multiple disparate noisy sensors Sejong Yoon, Mubbasir Kapadia, Pritish Sahu, Vladimir Pavlovic IEEE Winter Applications of Computer Vision Workshops (WACVW 2016) [pdf]

Patents

System and method for comprehension based question answering using taxonomy Ajay Divakaran, Michael Cogswell, Pritish Sahu US Patent App. 17/869,589, 2023 [pdf]

System and method for content comprehension and response Ajay Divakaran, Karan Sikka, Yi Yao, Yunye Gong, Stephanie Nunn, Pritish Sahu, Michael Cogswell, Jesse Hostetler, Sara Rutherford-quach US Patent App. 17/516,409, 2022 [pdf]

Dissertations

Cube maze Pritish Sahu, James Abello Master Dissertation [pdf]

Study of approaches to remove show-through and bleed-through in document images Pritish Sahu, Pankaj Kumar Sa Bachelor Dissertation [pdf]

Achievements

Piero Zamperoni Best Student Paper Award presented at ICPR 2022 for “DAReN: A Collaborative Approach Towards Visual Reasoning And Disentangling”. Featured in the IAPR newsletter, Volume 44, Number 4, October 2022 edition. [pdf]

Received “Outstanding Programming Application Award” and “Outstanding Project Award” from Computer Science Department at Rutgers University.

Received Samsung Best Project Award for successfully implementing Full HD & Smart TV features on Samsung TV.

Teaching Assistant

• CS440: Introduction to Artificial Intelligence, Summer 2019
  (Intrductory course on AI for the undergraduates)
• CS205: Discrete Structures I, Spring 2019
  (First course on discrete mathematics for the undergraduates)
• CS314: Principles of Programming Languages, Spring 2019
  (Fundamental course on programming languages for the undergraduates)
• CS206: Discrete Structures II, Fall 2018
  (Course on combinatorics and probability theory for the undergraduates)
• CS336: Principles of Information and Data Management, Spring, Summer 2018
  (an undergraduates course for describing and querying data such as in relational databases, XML, Web)
• CS526: Data Interaction and Visual Analytics, Spring 2018
  (an graduate course for big data algorithms, visualization and HCI)
• CS344: Desing and Analysis of Computer Algorithms, Fall 2017
  (an undergraduates course for algorithm design)

Projects

• Next Word Prediction using Long Short-Term Memory
  Built a language model trained using Recurrent Neural Network (LSTM) to predict the next word in a sentence using Theano. Compared with Bigram, Trigram and, 4-Gram Models to validate the belief that Recurrent Neural Network is better.
• Fast Trajectory Planning
  Implemented Adaptive A* search planning algorithm in Java/C++ that helps an agent move efficiently from a source to destination and adapt itself when an obstacle occurs in the path.[Code][Video]
• Face, Digit & Character Recognition
  Implemented character recognition by extracting the HU (centroid) moments from images and applied a Nearest Neighbor classifier to classify the test images. For face, digit classification implemented Perceptron, Naive Bayes & MIRA.
• Objects in Image Classification, Recognition & Detection
  The objective was to locate objects in shelf for the Amazon Pick up Challenge. The Challenge consisted in Object Classification, Detection, and Recognition. We used “SIFT & SVM” and “a Sequential Neural Network” using Theano and Keras for classification. Extracted Surf features to train and test for Detection. Trained on SIFT features and verified used homography and bounding box for Recognition.
• Web Portal for Master Students in CS (ms.cs.rutgers.edu)
  Full Stack portal for CS department with faculty and student information and interactions.
• Algorithm Snippets
  Application for tutoring algorithm via animation called “Algorithms Snippets”. It explains the Prims and Kruskal MST step by step.[Code][Video]

Misc

I enjoy playing badminton, swimming at my spare time.