PhD student
Computer Science Department
University of British Columbia
email: peymang@cs.ubc.ca
Dynamic Sparse Attention for Improving Efficiency in Foundation Models
Enhancing Localization and Image Understanding in Vision Language Models
Data Mixture
and
Agentic Skills
My research mainly deals with topics on HCI, AI, Machine learning.
o Principal Investigator of the Face Alignment, Eyelids, and Fatigue Metrics components of the Mirametrix Attention Sensing Engine
o Designed the structure and fined tuned various neural networks for facial landmark localization
o Mentored and defined projects for several interns working in machine learning, e.g., deep 3D face alignment for automatic labeling task
o Designed and built a ground truth system for the accurate evaluation of face
and head tracking systems
o Developed brand-new image processing and machine learning algorithms for the analysis of different medical/natural images
o Decoded big collection of medical data using batch scripts and constructed a platform for training deep neural networks for tracking retinal health
o Designed and implemented several visual perception experiments, e.g., relative judgment for autistic subjects, motion coherence, reverse phi phenomenon, etc.
o Formulated and introduced a brand-new active contour-based image segmentation algorithm
o Collected and set up an open access OCT image database
o Designed and prototyped a visual feedback for controlling bioprinter robot
o Proposed a computer vision protocol for wound depth estimation using disparity map
o Assisted on the installation, trouble shooting, and maintenance of different medical imaging devices, e.g. MRI, CT Scan, XRay at the customer service department.
o Administrated the safety standards of medical devices at the sales and marketing department
Monday, 14:30 - 15:30