I’m Arshiya Aggarwal, a master’s student in Computer Science at Columbia University. I’ve previously worked as a Software Development Engineer-2 at Adobe where I ideated, developed, optimized and deployed Machine Learning models for Adobe Scan. I graduated from Delhi Technological University in Electronics and Electrical Engineering. I have a keen interest in research in Natural Language Processing and am intrigued by the insights that can be gleaned from data.

Research

My research interests lie at the intersection of Deep Learning and Natural Language Processing. My goal is to build interpretable deep learning models that are more conducive to reducing societal bias. My research at MIDAS Lab was focused in NLP where I researched stock volatility prediction through multimodal analysis of CEOs’ speeches in financial Earnings Calls published papers at top conferences including EMNLP, Interspeech and NAACL.

News

• September 2021 : Started working as a Graduate TA/CA for COMS 4771 Machine Learning by Prof. Daniel Hsu
• September 2021 : Started pursuing master’s in Computer Science at Columbia University focussing my studies in Machine Learning and Natural Language Processing
• March 2021 : Poster on semantic segmentation of occlusions in scanned documents accepted at Adobe’s TechSummit 2021
• March 2021 : Paper on multimodal bias accepted at NAACL-HLT 2021
• January 2021 : Promoted to Software Development Engineer-2 at Adobe
• November 2020 : Presented long paper at EMNLP 2020
• November 2020 : Received honourable mention for 2 hacks at Adobe’s Hackweek 2020. One for book scan curvature removal and other for aspect-based sentiment analysis on Acrobat’s user review (Net Promoter Score) data
• Oct-Nov 2020 : Volunteered at EMNLP 2020 as a sponsor liaison to help set up virtual booths and the conference website
• October 2020 : Presented full paper at Interspeech 2020
• October 2020 : US Patent Ref# P9928-US filed as a part of collaborative research at Adobe titled Identity Obfuscation In Images Utilizing Synthesized Faces
• September 2020 : US Patent Ref# P9838-US filed as a part of collaborative research at Adobe titled Machine Learning Techniques For Differentibility Scoring Of Digital Images
• December 2019 : Best Project Award at Python training bootcamp at Adobe
• June 2019 : Graduated Delhi Technological University with department rank 2 in Electronics and Electrical Engineering