Ashim Gupta

I am a PhD student at the School of Computing, University of Utah advised by Prof. Vivek Srikumar. The main area of my research in NLP is the analysis, interpretability, and robustness of NLP models. I also frequently work on low-resource languages like Sanskrit focusing on incorporating Sanskrit specific lingusitc rules into structured prediction models for Sanskrit.

I am grateful to Bloomberg for supporting my research through the Bloomberg Data Science Ph.D. Fellowship.

Before this, I was a Research Assistant in the Computer Science Department at IIT Kharagpur. My research advisors at IIT Kharagpur were Prof. Pawan Goyal and Prof. Sudeshna Sarkar. My work at IIT Kharagpur mainly focussed on information extraction from biomedical literature and unsupervised machine translation systems for indic languages.

I completed my Bachelor’s Degree in Electrical Engineering from IIT-BHU in 2016. My undergraduate research advisors were Prof. Rajeev Srivastava and Prof. D. N. Vishwakarma. During my undergraduate, my research involved working on image retrieval systems both in general and medical domain. For my undergraduate thesis, I studied and implemented robust multi-sensor data fusion methods using Kalman Filter.

Email  /  CV  /  Google Scholar  /  LinkedIn

Publications

Whispers of Doubt Amidst Echoes of Triumph in NLP Robustness [paper]
Ashim Gupta, Rishanth Rajendhran, Nathan Stringham, Vivek Srikumar, Ana Marasović
arXiv Preprint

IntenDD: A Unified Contrastive Learning Approach for Intent Detection and Discovery [arxiv]
Bhavuk Singhal, Ashim Gupta, Shivasankaran V P, Amrith Krishna
Findings of EMNLP 2023

Don't Retrain, Just Rewrite: Countering Adversarial Perturbations by Rewriting Text [arxiv]
Ashim Gupta, Carter Wood Blum, Temma Choji, Yingjie Fei, Shalin Shah, Alakananda Vempala, Vivek Srikumar
ACL 2023

Adversarial Clean Label Backdoor Attacks and Defenses on Text Classification Systems [arxiv]
Ashim Gupta, Amrith Krishna
RepL4NLP @ ACL 2023

Sāmayik: A Benchmark and Dataset for English-Sanskrit Translation [arxiv]
Ayush Maheshwari, Ashim Gupta, Amrith Krishna, Ganesh Ramakrishnan, G. Anil Kumar, Jitin Singla
arXiv Preprint

X-FACT: A New Benchmark Dataset for Multilingual Fact Checking [arxiv][code][Talk][poster]
Ashim Gupta, Vivek Srikumar
ACL 2021

Evaluating Relaxations of Logic for Neural Networks: A Comprehensive Study [link]
Mattia Medina Grespan, Ashim Gupta, Vivek Srikumar
IJCAI 2021

BERT & Family Eat Word Salad: Experiments with Text Understanding [arxiv][Poster]
Ashim Gupta, Giorgi Kvernadze, Vivek Srikumar
AAAI 2021

A Little Pretraining Goes a Long Way: A Case Study on Dependency Parsing Task for Low-resource Morphologically Rich Languages [arxiv]
Jivnesh Sandhan, Amrith Krishna, Ashim Gupta, Laxmidhar Behera, Pawan Goyal
EACL-SRW, 2021

A Graph Based Framework for Structured Prediction Tasks in Sanskrit
Amrith Krishna, Ashim Gupta, Pawan Goyal, Bishal Santra, Pavankumar Satuluri
ACL Computational Linguistics Journal (December 2020 Issue)

Keep It Surprisingly Simple: A Simple First Order Graph Based Parsing Model for Joint Morphosyntactic Parsing in Sanskrit
Amrith Krishna, Ashim Gupta, Deepak Garasangi, Pavankumar Satuluri, Pawan Goyal
EMNLP 2020

Evaluating Neural Morphological Taggers for Sanskrit
Ashim Gupta, Amrith Krishna, Pawan Goyal, Oliver Hellwig
17th SIGMORPHON at ACL 2020

Neural Approaches for Data Driven Dependency Parsing in Sanskrit
Amrith Krishna, Ashim Gupta, Deepak Garasangi, Jivnesh Sandhan, Pavankumar Satuluri, Pawan Goyal
Technical Report (World Sanskrit Conference, 2021)

Fully Contextualized Biomedical NER
Ashim Gupta, Pawan Goyal, Sudeshna Sarkar
41st Eurpoean Conference on Information Retrieval, 2019

An LSTM-CRF Based Approach to Token-Level Metaphor Detection
Malay Pramanick, Ashim Gupta, Pabitra Mitra
NAACL Workshop on Figurative Language Processing, 2018

Content Based Mammogram Retrieval System
VP Singh*, Ashim Gupta*, Shubham Singh, Rajeev Srivastava
IEEE UPCON at IIIT Allahabad, 2015

Notes

Some useful lecture notes.

Maximum Likelihood Estimate for Multivariate Gaussian
It is known that for a multivariate Gaussian, maximum likelihood estimate for mean is unbiased and for variance, the estimate is biased. This document has the proof. [pdf]

Slides/Presentations

Generative AI and How Business Owners can use it
Slides from a presentation I gave at the Annual Veteran Business Conference with Maitrey on Generative AI and how business owners can use the latest AI tools. [pptx]


Credits to Jon Barron for this template.