Haocheng Dai
I am currently a research engineer at Meta. Previously, I was an applied scientist at Amazon working on Rufus.
I earned my Ph.D. in Computer Science from the University of Utah. I was fortunate to be mentored by Dr. Sarang Joshi. I also worked closely with researchers from FSU, UCLA, UVA, and Yale.
Before joining the University of Utah, I received my B.Eng. in Computer Science from Tongji University and studied at Israel Institute of Technology and Institut de Mathématiques de Toulouse as an exchange student, focusing on image analysis and Riemannian geometry, respectively.
Publications & Preprints
* = co-first author
The Silent Majority: Demystifying Memorization Effect in the Presence of Spurious Correlations.
Nature Communications, 5424 (2025).
Refining Skewed Perceptions in Vision-Language Models through Visual Representations.
ICCV Workshop on Multimodal Representation and Retrieval (MRR), 2025.
Applied Radiation Oncology (ARO), 2023.
Neural Operator Learning for Ultrasound Tomography Inversion.
International Conference on Medical Imaging with Deep Learning (MIDL), 2023.
Modeling the Shape of the Brain Connectome via Deep Neural Networks.
International Conference on Information Processing in Medical Imaging (IPMI), 2023. Oral Presentation.
Machine Learning for Biomedical Imaging (MELBA), 2022.
Structural Connectome Atlas Construction in the Space of Riemannian Metrics.
International Conference on Information Processing in Medical Imaging (IPMI), 2021. François Erbsmann Prize (Best Paper Award).
Services
I have served as a reviewer for several journals and conferences, including ACM MM, AISTATS, ACM TIST, CVPR, ARO, BMVC, ICLR, ICML, IEEE TNNLS, MedIA, MELBA, MICCAI, MIDL, NeurIPS, Scientific Reports, TMI, TMLR, AI for Differential Equations in Science@ICLR, MRR@ICCV and WiCV@ECCV.
Miscellaneous
I made a handful of notes for better understanding in language models, vision models, machine learning, mathematics of imaging, metric estimation, image registration, and solving large systems of linear equations.
My Erdős number = 4:
Haocheng Dai → Sarang Joshi → Ulf Grenander → Oved Shisha → Paul Erdős;
Haocheng Dai → Mike Kirby → Frank Stenger → Ambikeshwar Sharma → Paul Erdős.
I am an amateur photographer, vlogger and also a loyal reader of newspapers. You can find the highlight front pages of the New York Times I collect by the years of 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020, 2021, 2022, 2023, 2024, and 2025; the highlight front pages of the Washington Post I collect before 2015, 2016 - 2020, and 2021 - 2024.
I also write some random notes for fun.
Footprints
