I'm a CS PhD student at UNC Chapel Hill advised by Professor Henry Fuchs. My research interests are at the intersection of computer graphics, augmented reality, and virtual reality. I do research that typically explores these topics with an emphasis on having an outcome that directly benefits applications in areas such as healthcare, entertainment, and robotics.

I've done internships at Google and Kitware thus far in my PhD. Prior to pursuing a PhD, I was an embedded systems engineer working on wearable devices at Nike. I completed my undergraduate education in electrical and computer engineering at NC State University.

I'm happy to chat about research, new opportunities, and life in general with just about anyone. Feel free to reach out via email!

I'm actively looking for Summer 2025 internships in graphics, AR/VR, and application areas such as health and entertainment. Please reach out via email if you have an opportunity that that might be a strong fit!

News
  • Sep 2023 One first-author paper has been accepted to EMNLP 2024 (Main Track) in Miami, Florida!
  • Jul 2023 One first-author paper has been accepted to ECCV 2024 in Milan, Italy!
  • Mar 2023 Started an internship on Google's Consumer Health Research Team (CHRT) in Seattle, WA!
  • Mar 2023 New website! Stay tuned for more updates.
...show all...

Research


This website is a work in progress. Please see my Google Scholar for a full, more up-to-date list of my publications!

teaser img

What Are the Odds? Language Models Are Capable of Probabilistic Reasoning

Akshay Paruchuri, Jake Garrison, Shun Liao, John Hernandez, Jacob Sunshine, Tim Althoff, Xin Liu, Daniel McDuff

EMNLP 2024 (Main)

Probabilistic reasoning is a key challenge for large language models (LLMs). Our paper evaluates LLMs on three tasks of estimating percentiles, drawing samples, and calculating probabilties using real-world and idealized distributions. Techniques such as within-distribution anchoring, real-world context, and simplifying assumptions (e.g., Normal approximations) improved performance by up to 70%.

teaser img

Leveraging Near-Field Lighting for Monocular Depth Estimation from Endoscopy Videos

Akshay Paruchuri, Samuel Ehrenstein, Shuxian Wang, Inbar Fried, Stephen M. Pizer, Marc Niethammer, Roni Sengupta

ECCV 2024

Near-field lighting, emitted by the endoscope and reflected by the surface, is modeled as Per-Pixel Shading (PPS). PPS features are used to perform depth refinement on clinical data using teacher-student transfer learning and a PPS-informed self-supervision, ultimately achieving state-of-the-art results on colonoscopy data.

teaser img

Motion Matters: Neural Motion Transfer for Better Camera Physiological Measurement

Akshay Paruchuri, Xin Liu, Yulu Pan, Shwetak Patel, Daniel McDuff, Roni Sengupta

WACV 2024 (Oral)

Neural Motion Transfer serves as an effective data augmentation technique for PPG signal estimation from facial videos. We devise the best strategy to augment publicly available datasets with motion augmentation, improving up to 79% in inter-dataset testing involving five benchmark datasets and 47% over existing results using SOTA methods on PURE.



Software

teaser img

rPPG-Toolbox: Deep Remote PPG Toolbox

NeurIPS 2023 Datasets and Benchmarks Track

A comprehensive toolbox that contains unsupervised and supervised remote photoplethysmography (rPPG) models with support for public benchmark datasets, data augmentation, and systematic evaluation.