Xiaosong Yuan
Philadelphia,PA
Summary
Master student of University of Pennsylvania. Graduated from Zhejiang University international Campus, in which trained with UIUC's academic standard as well as all-English courses and obtained bachelor's degree from both universities. Hardworking with high GPA. Plentiful of team project experiences and was usually the leader. Excellent English communication and writing skills. Interested in machine learning and passionate about engaging in work that uses machine learning technologies to change the world.
Overview
2
2
years of professional experience
Work History
Intern
Shijiazhuang JingHua Electronic Coorporation
06.2022 - 08.2022
- Assisting project group to design a high resolution LED matrix with quantum Interference and diffraction, which is expected to have 54% light efficiency.
- Proposing 'A method based on a grating lens to solve the brightness uniformity of the display module light source', which solves the problem of excessive brightness differences between different pixels due to diffraction. This method reduces the brightness difference to within 20%, which is imperceptible to the human eye.
- The solution to the brightness uniformity has been patented by the company.
Senior Design
Farglory Life Insurance Company, Taiwan
03.2022 - 06.2022
- Senior design project at the International College of Zhejiang University, serving as the leader of a four-person team. Developed a high-precision handwritten Chinese character recognition system for Farglory Life Insurance Company, Taiwan.
- Expanded the database from the original 2,678,424 images with 3,755 Chinese character classes to 3,895,135 images with 7,356 classes. This includes all 6,763 commonly used Chinese characters specified in GB-2312, as well as numbers, English letters, and punctuation.
- Performed preprocessing on the new database. Tuned existing deep learning models and retrained them with the new database, achieving a test set accuracy of 88.74%.
- Wrote interface code, such as inputting handwritten text images cut from insurance contracts into the model and exporting the recognition results to the corresponding places in the Excel spreadsheet.
Summer Research Assistant
Professor Manolis Kellis From MIT
07.2021 - 08.2021
- Eight encoders were trained using eight databases based on Fabris's research on modular DNN to solve the problem of predicting genes of 27 aging related diseases. The contribution of each encoder to a accuracy was quantitatively studied.
- A new structure was proposed and a combiner was trained. The new model achieved 97.32% accuracy in predicting aging genes, which is a 10.65% improvement over the original paper.
- Paper: Xiaosong Yuan, Ruoyang Hong, and Danyating Shen. 2022. A Deeper Understanding of Modular DNN in Predicting Ageing-Related Disease. In 2021 4th International Conference on Digital Medicine and Image Processing (DMIP '21). Association for Computing Machinery, New York, NY, USA, 45–53. https://doi-org.proxy.library.upenn.edu/10.1145/3506651.3506659
Summer Research Assistant
Professor Fumiya Iida From Cambridge University
07.2020 - 08.2020
- Tested the convergence time of different parameters in the Bellman equation. Used Q-Learning algorithm with "Exploitation-Exploration Trade-Off" strategy and optimal parameters in the Robotics Playground simulation to plan a shortest path for a robot. The adoption of the optimal solution resulted has a 37% improvement in convergence speed compared to the original algorithm.
- Paper: Yuan X (2021) Faster Finding of Optimal Path in Robotics Playground Using Q-Learning with “Exploitation-Exploration Trade-Off”. J Phys Conf Ser 1748(2):022008
Education
Master of Science Engineering - Electrical Engineering
University of Pennsylvania
Philadelphia, PA05.2024
Bachelor of Science With Honor - Computer Engineering
University of Illinois At Urbana-Champaign
Champaign, IL06.2022
Bachalor of Engineering - Electronic And Computer Engineering
Zhejiang University
Zhejiang, China, 31440006.2022
Skills
- Machine learning, deep learning, AI, PyTorch/TensorFlow
- Parallel programming in CUDA
- Project, research experience and paper writing
- Programming languages including Python, C/C, SQL, Verilog, assembly
Timeline
Intern
Shijiazhuang JingHua Electronic Coorporation
06.2022 - 08.2022
Senior Design
Farglory Life Insurance Company, Taiwan
03.2022 - 06.2022
Summer Research Assistant
Professor Manolis Kellis From MIT
07.2021 - 08.2021
Summer Research Assistant
Professor Fumiya Iida From Cambridge University
07.2020 - 08.2020
Master of Science Engineering - Electrical Engineering
University of Pennsylvania
Bachelor of Science With Honor - Computer Engineering
University of Illinois At Urbana-Champaign
Bachalor of Engineering - Electronic And Computer Engineering
Zhejiang University
Similar Profiles
LIYANG ZHULIYANG ZHU
Drug Quality Inspection at Jinghua Pharmaceutical Group Co., LtdDrug Quality Inspection at Jinghua Pharmaceutical Group Co., Ltd
Haotian HuHaotian Hu
Human Resources Specialist (42A) at Sandy Industrial Co., Ltd., ShijiazhuangHuman Resources Specialist (42A) at Sandy Industrial Co., Ltd., Shijiazhuang
Jiaqi ZhenJiaqi Zhen
Salesperson at Shijiazhuang Shanjie Import&Export Co.,LtSalesperson at Shijiazhuang Shanjie Import&Export Co.,Lt