I am a Software Engineer at Google on the Pixel AI team, where I work on data and i18n to power on-device LLM features like Magic Cue. I hold a combined BS/MS in Computer Science from the University of California, Santa Barbara, where I was advised by Professor William Yang Wang in the NLP Group. My research focuses on assessing the reasoning capabilities of large language models and vision language models (LLMs & VLMs) — you can read more in my thesis.

News

[Dec 2024] T2IScoreScore presented as a Spotlight Paper at NeurIPS 2024!

[Nov 2024] EMNLP 2024 Poster Presentation Session in Miami, Florida, USA!

[Jun 2024] Defended my Master's thesis with flying colors!

[Jan 2024] Presented Special Session Oral paper OCTO+ at the IEEE AIxVR Conference in USC Institute of Creative Technologies!

[Jun 2023] Software Engineering Intern at Google working on Gmail, Cloud Workspace team.

[Jun 2023] Invited to speak at 2023 UC Santa Barbara Foundation Board of Trustees meeting!

[May 2023] ICLR 2023 Oral Presentation at Kigali Convention Centre in Kigali, Rwanda.

[Jan 2023] Extremely excited to share that WikiWhy has been accepted as an Oral Paper: Top 5% out of all 4019 submissions at ICLR 2023!

Research

Losing Visual Needles in Image Haystacks: Vision Language Models are Easily Distracted in Short and Long Contexts
Aditya Sharma*, Michael Saxon*, William Yang Wang
Findings of Empirical Methods in Natural Language Processing (EMNLP), 2024
Paper Homepage

Who Evaluates the Evaluations? Objectively Scoring Text-to-Image Prompt Coherence Metrics with T2IScoreScore (TS2)
Michael Saxon*, Fatima Jahara*, Mahsa Khoshnoodi*, Yujie Lu, Aditya Sharma, William Y. Wang
38th Conference on Neural Information Processing Systems (NeurIPS), 2024
Paper Homepage

OCTO+: A Suite for Automatic Open-Vocabulary Object Placement in Mixed Reality
Aditya Sharma*, Luke Yoffe*, Tobias Höllerer
IEEE International Conference on Artificial Intelligence and eXtended and Virtual Reality (AIxVR)
Paper Code Slides Homepage

OCTOPUS: Open-vocabulary Content Tracking and Object Placement Using Semantic Understanding in Mixed Reality
Aditya Sharma*, Luke Yoffe*, Tobias Höllerer
IEEE International Symposium on Mixed and Augmented Reality Adjunct (ISMAR-Adjunct), 2023
Paper Code

WikiWhy: Answering and Explaining Cause-and-Effect Questions
Aditya Sharma*, Matthew Ho*, Justin Chang*, Michael Saxon, Sharon Levy, William Yang Wang
International Conference on Learning Representations (ICLR), 2023 [Top 5% Oral Paper]
Paper Code Slides

Pin&Post — App-Based Reporting of Electrical Fire Hazards to Prevent Wildfires
Aditya Sharma, Ishan Goyal, Silvia Figueira
IEEE Global Humanitarian Technology Conference (GHTC), 2020
Paper