Gargi Balasubramaniam

Selected Papers | Selected Projects | Selected Talks | News | Contact

I am a 2nd year MS CS student at the University of Illinois, Urbana Champaign (UIUC). I am fortunate to be advised by Prof. Han Zhao in the field of transfer learning. Specifically, I am working on domain generalization and multimodal machine learning from the lens of robustness and interpretability.

I graduated as a Gold Medallist from BITS Pilani, Goa with a Bachelor’s in Computer Science, where I worked with Prof. Ashwin Srinivasan. Prior to joining UIUC, I have had wonderful experiences of working at Microsoft (Software Engineer), American Express (ML Research Intern), and at the Techincal University of Munich Germany (DAAD Scholar).

Currently, I am interning at Meta Reality labs as a SWE intern. If you are in the bay area and would like to catch up, please feel free to drop a message !

🎵🎵 I am a music buff! A trained Hindustani Classical Vocalist, and like to play the guitar,piano and compose tunes. You will often find me at the piano situated at Illini Union, UIUC . Checkout my covers here!

Affiliations


BITS Pilani, Goa Gold Medallist 2016-2020	Microsoft 2020-2021	UIUC 2021-Present

Internships


TIFR	Technical University of Munich	American Express	Meta
Summer 2018	Summer 2019	Spring 2020	Summer 2022
SWE Intern	DAAD Scholar	ML Research Intern	Reality Labs, SWE Intern
e-learning systems	Interpretable ML	Model Reliability & Callibration	Augmented Reality

selected papers Full List

Greedy Modality Selection via Approximate Submodular Maximization

Gargi Balasubramaniam*, Runxiang Cheng*, Yifei He*, Yao-Hung Hubert Tsai, and Han Zhao

In UAI 2022, The 38th Conference on Uncertainty in Artificial Intelligence 2022

Abs

Camera Ready Coming Soon!
When does Worst Group Performance Signify Robustness to Spurious Features?

Gargi Balasubramaniam

2022

Abs

Article Coming soon.
Towards Comprehensible Representation of Controllers Using Machine Learning

Gargi Balasubramaniam

In ASE 2019, The 34th IEEE/ACM International Conference on Automated Software Engineering 2019

Abs Link

From the point of view of a software engineer, having safe and optimal controllers for real life systems like cyber physical systems is a crucial requirement before deployment. Given the mathematical model of these systems along with their specifications, model checkers can be used to synthesize controllers for them.The given work proposes novel approaches for making controller analysis easier by using machine learning to represent the controllers synthesized by model checkers in a succinct manner, while also incorporating the domain knowledge of the system. It also proposes the implementation of a visualization tool which will be integrated into existing model checkers. A lucid controller representation along with a tool to visualize it will help the software engineer debug and monitor the system much more efficiently.
Use of Machine Learning on Channel State Information in IoT-based WiFi Networks for Movement Detection

Y. Dasgaonkar, G. Balasubramaniam, and V. Naik

In COMSNETS 2020, The 12th International Conference on Communication Systems & Networks 2020

Abs Link

With the ubiquitous nature of WiFi devices and WiFi networks around us, which continuously transmit and receive high bandwidth data, these transmission channels can be used as an effective sensing medium for detecting movements. With the use of Machine Learning, we develop a model using the features of the channel state information to train and predict these movements accurately. This paper proposes a demonstration harnessing the low-cost WiFi-based IoT for the detection of movements. Such a setup will facilitate the usage of movements like hand gestures to carry out common tasks such as changing the television channel, changing slides during a presentation, no-controller wireless gaming, etc. We achieve an F1 score of 0.99 with the latency of 45.75ms.

selected projects Full List

An empirical analysis of Invariant Representations for Domain Generalization in Vision

Computer Vision, UIUC, Spring 2022

Abs

We present an empirical analysis of different domain generalization algorithms - a setting where the learner has access to several domains, and can leverage these data points to best generalize to new, unseen domains. In this work, we present a structured approach to compare and contrast representations learnt by these methods, both qualitatively and quantitatively. We enable this analysis by treating each algorithm as a descriptor producing oracle. Our work is evaluated on the PACS dataset - containing images in the styles of Photos, Art Paintings, Cartoon and Sketch. We propose the following metrics and their motivations: a) Test-time performance and fairness evaluated by accuracy at 3 levels of granularity. b) Interpretability and Reliability via calibration Analysis c) Domain Invariance via qualitative TSNE Visualization.
Learning Descriptors for Image Matching and Copy Detection

Computer Vision, UIUC, Fall 2021

Abs

Image similarity estimation and copy detection are upcoming research areas. Given today’s era of social interaction - social media serves as a ground for spread of misinformation, with manipulated images being a common modality. In order to tackle this issue, image matching is a key component of content tracing systems which are used to flag manipulated and violating content. Furthermore, a wide variety of trans- formations are possible by using advanced and accessible image editing software - which implies that this is not a trivial problem. In this context, we aim to solve the following problem: can we determine whether an image is an edited copy of another image? We take inspiration from the Image Similarity Challenge 2021conducted at CVPR ’21 and NeurIPS ’21. Specifically, given a query image (which may be manipulated) - we aim to develop a model that can predict whether this image is a derived from a large set of reference images, and find the corresponding matching image from this reference set. Our approach to achieve this goal is to learn reliable feature descriptors for each image, which can be used to obtain the most similar images from the reference set.

selected presentations Full List

Paper Review: "Fine-Tuning can Distort Pretrained Features and Underperform Out-of-Distribution", Ananya Kumar et al., ICLR 2022

TML Reading Group, UIUC, April 2022

Slides Original Paper
Paper Review: "The pitfalls of simplicity bias in neural networks", Harshay Shah et al., NeurIPS 2021

TML Reading Group, UIUC, March 2022

Slides Original Paper
Paper Review: "Impossibility Theorems for Domain Adaptation", Ben David, et al., JMLR 2010

Transfer Learning, UIUC (Co-presented with Sam Cheng, Yifei He), Nov 2021

Slides Original Paper

news

May '22	I started my internship at Meta Reality Labs as a SWE intern. If you’re in the Bay Area, lets connect!
Apr '22	[UAI 2022] Our paper on modality selection via submodular maximization is accepted at UAI 2022!
Aug '21	I have joined the MS CS program at UIUC! Looking forward to a wonderful time 🌽
Jul '21	Organizing a session on Variational Inference in NNs as part of the WiML UnWorkshop at ICML 2021!
Mar '21	Serving as a reviewer on the ICLR 2021 Workshop on Rethinking ML papers!
Nov '20	Participated in interesting conversations as part of the Google PAIR Symposium 2020
Nov '20	I will be volunteering for NeurIPS 2020!
Nov '20	I have been selected to attend the DS4SG Workshop at Georgia Tech! Key insights coming soon.
Sep '20	I participated in a closed QnA with Sir Tony Hoare and Leslie Lamport as part of vHLF 2020.
Sep '20	Honoured to be the Gold Medalist, Graduating batch of 2020, BITS Pilani, Goa.
Aug '20	Thrilled to be a panelist for the "Research Saturday" webinar on Research Internships by CLIMB DTU
Jul '20	Excited to attend my first ever ICML conference!
Nov '19	Awarded the third position for the ACM SRC (Undergraduate) at ASE 2019
Nov '19	Awarded the SIGSOFT CAPS Travel Grant for attending ASE 2019.
Sep '19	Stood 5th at the Singapore India Hackathon 2019!
Aug '19	[COMSNETS 2020] Demonstration paper accepted at COMSNETS 2020!
Aug '19	[ASE 2019]Paper accepted at ASE 2019,San Diego for the SRC track!
May '19	I will be spending my summer as a DAAD Scholar at the Technical University of Munich, Germany.
Apr '19	Won the Smart India Hackathon 2019 for a problem statement by ISRO!
Mar '19	Selected for the MITACS Globalink Research Scholarship and the DAAD WISE Scholarship