Chang Liu
Development
PA, United States
Skills
Python
About
Chang Liu's skills align with IT R&D Professionals (Information and Communication Technology). Chang also has skills associated with Consultants and Specialists (Information and Communication Technology). Chang Liu appears to be an entry-level candidate, with 4 months of experience.
View more
Work Experience
Research Scientist intern
Whiterabbit.ai
June 2022 - August 2022
- Refined breast cancer early detection algorithm for improved performance Developed and fine-tuned a novel model by combining self-attention and convolutional layers, leading to significantly improved performance for classification Employed spatial transformer block to address geometric variations in input images Conducted comprehensive evaluation of model performance using a large medical image dataset consisting of ~150,000 images Implemented and conducted comparative analysis between Sharpness-Aware Minimization (SAM) and Adaptive Sharpness-Aware Minimization (ASAM) to optimize model training
Al/ML Ph.D. intern
GE Healthcare
June 2021 - August 2021
- Optimized 3D deep learning segmentation model for CT radiotherapy planning Employed data-driven data augmentation to train the model in a data-limited setting, maintained similar performance using only 5% of data Managed cloud instances with SageMaker and E3 on AWS infrastructure, facilitating the deployment and scalability of large-scale deep-learning segmentation models Projects Built interpretable AI models for cardiac arrest outcome prediction with head CT images Designed a multimodal machine learning model for clinical outcome prediction Link to article Identified clinical-relevant image patterns for clinical outcome prediction with interpretable machine learning model Link to article Designed a new deep learning algorithm for imbalanced medical image data classification Developed a novel self-supervised learning-inspired one-class classification model to optimally learn single-class-relevant inherent imaging features Link to article Applied deep learning model to predict human induced Pluripotent Stem Cell (hiPSC) differentiation Conducted a comparative analysis of Convolutional Neural Network (CNN) and transformer models to assess their performances in predicting cell function
Education
Georgia Institute of Technology
MS
Beihang University
BS
