Hi, I'm Ruchira Ray
(she/her)

I am a second-year master's student in the Department of Computer Science at UT Austin advised by Dr. Amir Zamir and Dr. Philipp Krähenbühl where I am working on controlled data generation. I am also a member of the Robot Interactive Intelligence Lab, where I am researching the motivation behind the desire for robotic automation of household tasks and its social effects as a part of the UT Austin Good Systems.

Previously, as at Samsung, I worked on encoder-decoder models for audio anti-spoofing. I have also worked on mathematical equation recognition (as a MITACS Globalink Intern), and on autoencoder models for recommender systems, under the supervision of Dr. Bidyut Kumar Patra.

Click here to know more!

Email  /  CV  /  Github  /  LinkedIn  /  Google Scholar  /  Twitter

I am broadly interested in machine learning models that facilitate collaboration between humans and machines. I'm interested in multimodal learning for visual perception and robotics. I'm also interested in incorporating them for social good.
I find probabilistic and variational ML models to be very interesting.

• June 2023: Won Best Paper Award at ACM FAccT23 for Queer In AI: A Case Study in Community-Led Participatory AI
• December 2022: Organised Queer in AI Affinity Workshop Neurips 2022.
• March 2022: Major Project "Cost-efficient Gait Recognition using 3D CNN" selected as one of the top 10 in the Computer Science and Engineering Department.
• December 2021: Organised Queer in AI Affinity Workshop Neurips 2021.
• July 2021: Presented our work "Feature Genuinization based Residual Squeeze-and-Excitation for Audio Anti-Spoofing in Sound AI" at 12th ICCCNT, 2021. Presented poster on the same at the "Queer in AI" workshop, ICML'21
• July 2021: Attended Eastern European Machine Learning Summer School
• January 2021: Selected to attend Mediterranean Machine Learning (M2L) Summer School and presented our poster DBVAE: Deep Belief Variational Autoencoder for top-N Recommender Systems
• February 2019: Won a silver medal at SRM University Research Day for my paper presentation on music genre classification
• June 2018: Won SRMJEE Merit Scholarship from SRM University that waives 50 percent of my tuition

* = Equal Contribution

Method to control text-to-image generation system using guided adversarial prompts for effective training data generation.

Queer in AI as a case study for community-led participatory design in AI

Squeeze-and-Excitation encoder decoder model for fake/spoofed audio detection

Deep Belief Network and Variational Autoencoder hybrid architecture for collaborative filtering

Summarizing and comparing different data augmentation strategies using S3PRL toolkits in speech-processing tasks.

Multiple prompting strategies to jailbreak LLMs' decision-making processes and help them balance refusing unsafe prompts without decreasing helpfulness.

Foster awareness of LGBTQ+ issues, cultivate a dynamic community of LGBTQ+ researchers and illuminate the remarkable contributions of LGBTQ+ scientists in the field of robotics. Currently organising Welcome Social for IROS’23.

Actively working towards increasing the participation of LGBTQ+ individuals in AI and spreading awareness of diversity in STEM. Currently organising QAI Workshop for Neurips’23.

Spearhead a team of 45+ students to organize events, increase outreach and supervise projects. Outreach and number of events tripled in the year 2020-2021.

Our lab won the Silver prize for Student-led Innovation award (2018) by QS Reimagine Education and the Wharton School of Business.

Working on multidisciplinary pet-projects, incharge of mentoring associates, conducting interviews for new recruits , managing events etc.