Gaurav Kumar Nayak

Email: / Google scholar / Semantic scholar / Linkedin

Currently, doing postdoc at University of Central Florida under Dr. Mubarak Shah . I completed my PhD at Visual Computing Lab, CDS, Indian Institute of Science, under the supervision of Prof. Anirban Chakraborty. Prior to that I completed my M.Tech from Jawarharlal Nehru University under the guidance of Prof. R. K. Agrawal, and B.Tech from VIT University in computer science. My area of research focuses on Artificial Intelligence at the intersection of Computer Vision, Machine Learning and Deep Learning.


The performance of any deep learning model depends heavily on the data, using which it has been trained. The quality of the trained deep models improve with the availability of large number of uncorrelated datapoints. If the data available for training is not sufficiently large, the deep models overfit and do not generalize. Evidently, data plays a very crucial role in training any deep neural network model. However, data is ‘precious’, and may not be freely available all the time. Many companies have proprietary rights over their data. For example, models trained by Google and Facebook might utilize proprietary data such as JFT-300M, SFC. Also, the data may not be shared if it contains sensitive information, specially, when dealing with biometric data of large population, healthcare data of patients etc. In a nutshell, data is often more precious than anything else in the era of deep learning and hence efficient utilization of data is of utmost importance. My research work in my PhD, therefore, looked into developing data efficient deep learning algorithms for computer vision tasks and was supported by fellowship from UGC-NET (JRF).



GeoCLIP: Clip-Inspired Alignment between Locations and Images for Effective Worldwide Geo-localization

Vicente Vivanco Cepeda, Gaurav Kumar Nayak, Mubarak Shah

In NeurIPS 2023

[Arxiv] [Slides

DAD++: Improved Data-free Test Time Adversarial Defense 

Gaurav Kumar Nayak, Inder Khatri, Shubham Randive, Ruchit Rawal, Anirban Chakraborty

In IJCV 2023 (Under Review)


DISBELIEVE: Distance Between Client Models is Very Essential for Effective Local Model Poisoning Attacks

Indu Joshi, Priyank Upadhya, Gaurav Kumar Nayak, Peter Schüffler, Nassir Navab 

In DeCaF, MICCAI Workshop 2023


Federated Learning on Heterogeneous Data via Adaptive Self-Distillation 

M Yashwanth, Gaurav Kumar Nayak, Arya Singh, Yogesh Simmhan, Anirban Chakraborty

In TNNLS 2023 (Under Review)


Data-efficient Deep Learning Algorithms for Computer Vision Applications

Gaurav Kumar Nayak

PhD Thesis

[Thesis] [Pdf] [Slides]

Data-free Defense of Black Box Models Against Adversarial Attacks

Gaurav Kumar Nayak, Inder Khatri, Shubham Randive, Ruchit Rawal, Anirban Chakraborty

In TIFS 2023 (Under Review)

[Arxiv] [Slides] [Code]

Robust Few-shot Learning Without Using any Adversarial Samples

Gaurav Kumar Nayak, Ruchit Rawal, Inder Khatri, Anirban Chakraborty

In TNNLS 2023 (Under Review)

[Arxiv] [Slides] [Code]

DE-CROP: Data-efficient Certified Robustness for Pretrained Classifiers

Gaurav Kumar Nayak*, Ruchit Rawal*, Anirban Chakraborty

In WACV 2023

[Arxiv] [Slides] [Poster] [Video] [Project Page

Holistic Approach to Measure Sample-level Adversarial Vulnerability and its Utility in Building Trustworthy Systems

Gaurav Kumar Nayak*, Ruchit Rawal*, Rohit Lal*, Himanshu Patil, Anirban Chakraborty

In HCIS, CVPR Workshop 2022

[Paper] [Arxiv] [Slides] [Poster] [Project Page]

DAD: Data-free Adversarial Defense at Test Time

Gaurav Kumar Nayak*, Ruchit Rawal*, Anirban Chakraborty

In WACV 2022

[Paper] [Supplementary] [Arxiv] [Slides] [Poster] [Video] [Project Page

Beyond Classification: Knowledge Distillation using Multi-Object Impressions

Gaurav Kumar Nayak*, Monish Keswani*, Sharan Seshadri, Anirban Chakraborty

In BMVC 2021

[Paper] [Supplementary] [Arxiv] [Slides] [Project Page

Incremental Learning for Animal Pose Estimation using RBF k-DPP

Gaurav Kumar Nayak*, Het Shah*, Anirban Chakraborty

In BMVC 2021

[Paper] [Supplementary] [Arxiv] [Slides] [Project Page

Mining Data Impressions from Deep Models as Substitute for the Unavailable Training Data

Gaurav Kumar Nayak, Konda Reddy Mopuri, Saksham Jain, Anirban Chakraborty

In TPAMI 2021

[Paper]  [Supplementary]  [Arxiv

Effectiveness of Arbitrary Transfer Sets for Data-free Knowledge Distillation

Gaurav Kumar Nayak*, Konda Reddy Mopuri*, Anirban Chakraborty

In WACV 2021

[Paper]  [Supplementary]  [Slides

DeGAN: Data-enriching GAN for retrieving representative samples from a trained classifier

Sravanti Addepalli, Gaurav Kumar Nayak, Anirban Chakraborty, R. Venkatesh Babu

In AAAI 2020

[Paper]  [Arxiv]  [Code] [Slides] [Poster]

Fusion of Deep and Non-Deep Methods for Fast Super-Resolution of Satellite Images

Gaurav Kumar Nayak, Saksham Jain, R. Venkatesh Babu, Anirban Chakraborty

In IEEE BigMM 2020

[Paper]  [Arxiv]  [Slides

Zero-Shot Knowledge Distillation in Deep Networks

Gaurav Kumar Nayak*, Konda Reddy Mopuri*, Vaisakh Shaj*, R. Venkatesh Babu, Anirban Chakraborty

In ICML 2019

[Paper]  [Supplementary]  [Code]  [Slides]  [Poster]

Efficient Person Re-identification in videos using Sequence Lazy Greedy Determinantal Point Process (SLGDPP)

Gaurav Kumar Nayak, Utkarsh Shreemali, R. Venkatesh Babu, Anirban Chakraborty

In ICIP 2019

[Paper]  [Slides]  [Poster]

Classification of Normal Versus Malignant Cells in B-ALL White Blood Cancer Microscopic Images

Atmika Honnalgere, Gaurav Kumar Nayak

In ISBI 2019

[Paper]  [Slides]  

Development and Comparative Analysis of Fuzzy Inference Systems for Predicting Customer Buying Behaviour

Gaurav Kumar Nayak, Swathi J Narayanan, Ilango Paramasivam

In IJET 2013


Data Analysis and Visualization 

Spatio-Temporal Analysis on NASA satellite data over Indian Region [Web Link