Shishira R Maiya

I am a PhD Student at University of Maryland, under Prof Abhinav Shrivastava where I work on computer vision and machine learning. Prior to this I earned my Masters in Computer Science from University of Maryland as well.

Previously, I finished my undergrad from MSRIT Bangalore. My first brush with machine learning was as an intern at Stride where I worked on text classification and other NLP tasks. I later went on to work under Prof AG Ramakrishnan on super resolution of document images at MILE lab at IISc. I continued working in related domains with an internship at Hyperverge under Adarsh Tadimari, where I worked on scene text recognition and OCR.

After my undergrad, I worked as a research assistant at RBCCPS at IISc, under Prof Raghu Krishnapuram. Here, I worked on a wide variety of problems ranging from multi-camera multi-object tracking, segmentation and vision guidance for drones. During my Masters, I had the opportunity of working as a research intern for SIML team at Apple, under Rui Shen.

I recently worked as a research intern at Snap Inc, where I worked on developing an efficient and scalable implicit neural representation for videos.

Currently my research spans model, data compression techniques and understanding adversarial examples for vision tasks.

Email  /  CV  /  Google Scholar  /  Github  /  Substack

profile photo
Research
Latent-INR: A Flexible Framework for Implicit Representations of Videos with Discriminative Semantics
Shishira R Maiya*, Anubhav Gupta* , Matt Gwilliam , Max Ehrlich , Abhinav Shrivastava
Under Review , 2024

We can now use INRs for downstream tasks !!

Project Page | Paper
NIRVANA: Neural Implicit Representations of Videos with Adaptive Networks and Autoregressive Patch-wise Modeling
Shishira R Maiya*, Sharath Girish* , Max Ehrlich , Hanyu Wang, Kwot Sin Lee, Patrick Poirson Pengxiang Wu Chen Wang, Abhinav Shrivastava
IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2023

An efficient and scalable INR for Videos!

Project Page | Paper
A Frequency Perspective of Adversarial Robustness
Shishira R Maiya, Max Ehrlich , Vatsal Agarwal, Ser-Nam Lim, Tom Goldstein, Abhinav Shrivastava
Under Review , 2021

A study of adversarial robustness from the perspective of frequency analysis.

Project Page | Paper
The Lottery Ticket Hypothesis for Object Recognition
Sharath Girish* , Shishira R Maiya*, Kamal Gupta , Hao Chen, Larry Davis, Abhinav Shrivastava
IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2021

Explored the nuances of applying the Lottery ticket Hypothesis for Object recognition models.

Project Page | Paper | Code
Membership Inference Attacks on Lottery Ticket Networks
Aadesh Bagmar* , Shishira R Maiya*, Shruti Bidwalkar , Amol Deshpande
International Conference on Machine Learning (ICML) workshop: A Blessing in Disguise: The Prospects and Perils of Adversarial Machine Learning , 2021

Are lottery ticket networks vulnerable to Membership inference attacks due to their sparse nature?

Paper | Code
Rethinking Retinal Landmark Localization As Pose Estimation: Naive Single Stacked Network For Optic Disk And Fovea Detection
Shishira R Maiya*, Puneet Mathur*
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2020

Can we use pose models for medical imaging problems ?

Paper
Slum Segmentation and Change Detection: A Deep Learning Approach
Shishira R Maiya*, Sudharshan Chandra Babu*
Neural Information Processing Systems (NeurIPS): ML4D workshop , 2018

Detecting and tracking growth of slums in the city of Mumbai using deep learning.

Project Page | Paper | Code
Relation Networks for Optic Disc and Fovea Localization in Retinal Images
Sudharshan Chandra Babu* , Shishira R Maiya*
Neural Information Processing Systems (NeurIPS): ML4Health workshop , 2018

Utilizing context for detection in medical imaging.

Paper
A new approach for upscaling document images for improving their quality
Rama Krishna Pandey, Shishira R Maiya, A G Ramakrishnan
IEEE India Conference (INDICON) , 2017

Method to upsample Document Images.

Paper

Template credits