username

Aditya Chinchure

Multimodal Learning @ UBC | Photography

👋 Hello there!

I am a PhD student at the Computer Vision and NLP Labs at University of British Columbia. I am fortunate to be advised by Dr. Leonid Sigal and Dr. Vered Shwartz, and supported by the highly competitive UBC Four Year Fellowship (4YF).

My research focus is on applying methods from human reasoning and cognition to develop vision-language models 👁️ and evaluation tools 🛠️. I have worked on image and video understanding & reasoning, as well as bias and fairness in Generative AI.

I recently completed my internship at Ideogram, where I worked on Ideogram 4.0. Previously, I was a visiting scholar at Toyota Technological Institute at Chicago, working with Dr. Matthew Turk, and an intern at Borealis AI with Dr. Fred Tung. In addition, I enjoy doing photography as a hobby. My work has over 200 million views on Unsplash. You can find my latest photos on Instagram.

If you are reading this, I would love to talk to you! I am always looking for opportunities to collaborate. Also, my inbox is open if you have any questions about student life at UBC or Vancouver. Message me on Instagram or send me an email.

Google Scholar • Resume • Projects • LinkedIn

🗞️ News

📣 See More

📚 Publications

Spotlight: Identifying and Localizing Video Generation Errors Using VLMs
Aditya Chinchure, Sahithya Ravi, Pushkar Shukla, Vered Shwartz, Leonid Sigal
Accepted at ECCV 2026
arXiv

SPIKE-RL: Video-LLMs meet Bayesian Surprise
Sahithya Ravi, Aditya Chinchure, Raymond Ng, Leonid Sigal, Vered Shwartz
Accepted at ICLR 2026
arXiv

Position: World Models must live in Parallel Worlds
Sahithya Ravi*, Aditya Chinchure*, Pushkar Shukla, Vered Shwartz, Leonid Sigal (* equal)
Accepted at NeurIPS 2025 Workshop on Bridging Language, Agent, and World Models for Reasoning and Planning (LAW)
OpenReview | Paper

Mitigate One, Skew Another? Tackling Intersectional Biases in Text-to-Image Models
Pushkar Shukla*, Aditya Chinchure*, Emily Diana, Alexander Tolbert, Kartik Hosanagar, Vineeth Balasubramanian, Leonid Sigal, Matthew Turk (* equal)
Accepted at EMNLP 2025 (Findings)
arXiv | ACL Anthology

Black Swan: Abductive and Defeasible Video Reasoning in Unpredictable Events
Aditya Chinchure*, Sahithya Ravi*, Raymond Ng, Vered Shwartz, Boyang Li, Leonid Sigal (* equal)
Accepted at CVPR 2025
arXiv | Website

From Local Concepts to Universals: Evaluating the Multicultural Understanding of Vision-Language Models
Mehar Bhatia, Sahithya Ravi*, Aditya Chinchure*, Eunjeong Hwang, Vered Shwartz (* equal)
Accepted at EMNLP 2024
arXiv | Website

TIBET: Identifying and Evaluating Biases in Text-to-Image Generative Models
Aditya Chinchure*, Pushkar Shukla*, Gaurav Bhatt, Kiri Salij, Kartik Hosanagar, Leonid Sigal, Matthew Turk (* equal)
Accepted at ECCV 2024
arXiv | Website

Visual Question Answering with Contextualized Commonsense Knowledge [Masters Thesis]
Aditya Chinchure
UBC Library

VLC-BERT: Visual Question Answering with Contextualized Commonsense Knowledge
Sahithya Ravi*, Aditya Chinchure*, Leonid Sigal, Renjie Liao, Vered Shwartz (* equal)
Accepted at WACV 2023
arXiv | Code

Refinement Architectures for Referring Image Segmentation [Honours Thesis]
Aditya Chinchure
Thesis

LEAP: Private and Federated Data Analysis for Healthcare
Matheus Stolet, Chris Yoon, Kalli Leung, Aditya Chinchure, Mathias Lécuyer, Aline Talhouk, Ivan Beschastnikh
Poster at Emerging Technologies: BC’s AI Showcase, organized by UBC’s Centre for Artificial Intelligence Decision-making and Action (CAIDA)
Website

🌎 Collaborators

“It takes a village to raise a child, and a brilliant team (and lots of caffine) to raise a PhD”

Apart from my supervisors, I have had the pleasure of collaborating with:

… and many more.

📑 Reviewing

I have reviewed several papers for: NeurIPS 2025, CVPR 2025 (⭐️ Outstanding Reviewer), ICCV 2025, TPAMI 2024, ECCV 2024 (⭐️ Outstanding Reviewer), CVPR 2024, TPAMI 2023, TPAMI 2022

👨‍💻 Work

PhD Student, CV & NLP at UBC
Vancouver | May 2024 - Present
Building reasoning-inspired methods to evaluate and improve vision-language and generative AI models: fine-grained image and video understanding & evaluation, surprise localization and error detection in videos, and improving the diversity and quality of image and video generation.

Machine Learning Research Intern, Ideogram
Toronto | November 2025 - March 2026
Core research team member for Ideogram 4.0, ranked #1 open-weight model on DesignArena, with native layout specification. Designed 3+ VLM-based metrics and agentic pipelines for fine-grained text-to-image evaluation, RL, and system prompt optimization.

Visiting Student Researcher, Toyota Technological Institute at Chicago
Chicago | October 2024 - December 2024
Built InterMit, a counterfactual method for intersectional bias mitigation in image generation, achieving lower bias and higher diversity in fewer steps, with better image quality than prior work.

Machine Learning Research Intern, Borealis AI, RBC
Vancouver | September 2022 - May 2023
Developed PD-EST, a process-disentangling transformer for event sequences such as financial transactions. It jointly learns a process mask and the sequence model, improving next-event time/type prediction with interpretable process decomposition.

Graduate Research Assistant, Computer Vision Lab at UBC Computer Science
Vancouver | May 2022 - April 2024
Visual Question Answering with external commonsense knowledge.

Undergraduate Research Assistant, Computer Vision Lab at UBC
Vancouver | May 2020 - August 2020
Worked on implementing and evaluating structured attention for vision-text transformer models to improve image grounding.

Undergraduate Research Assistant, LEAP Project, UBC & BC Cancer Research
Vancouver | May 2020 - August 2020
Designed a module that enabled 2x faster data retrieval from REDCap research databases for LEAP, a privacy-focused federated ML platform.

Machine Learning Engineer (Co-op), Hypercontext (prev. SoapBox)
Toronto | May 2019 - August 2019
Developed BERT models for text classification, sentiment analysis and entity recognition, used in the Meeting Insights feature of the product. Orchestrated an end-to-end pipeline including data cleanup, a Flask REST API for serving, and retraining and deployment with Docker on AWS.

Junior Software Developer (Co-op), Broadcom (prev. AppNeta)
Vancouver | September 2018 – April 2019
Worked on scaling up our AppNeta’s platform for network monitoring for cloud deployments.