About Me
I am currently a second year Master student at Computer Science department, Sichuan University. My primary research interests lies in both practical and theoretical aspects of AI and machine learning. To elaborate, I enjoy making related algorithms more robust, efficient, explainable and powerful.
I like swimming and l especially enjoy the tranquil feeling in the water. l am also a fan of pop songs, computer games and photography.
- Name: Xuanzhi Feng
- Phone: +86 - 139 - 8110 - 6646
- Address: Chengdu, China
- Email: fengxuanzhi34@gmail.com
- Freelance: Available
Skills
%
Programming: Node.js, Python, JAVA, OCaml, LaTeX, C++, golang, Matlab, R, Stata
%
Devops: AWS, Docker, Kubernetes, Rancher, Vagrant, Packer, Terraform, Jenkins, CircleCI
%
Model:Transformer, Adversarial Learning, Meta Learning, Causal interference, incremental classifier and representation learning model and basic financial engineering models.
%
Languages: Fluent in English and Native in Chinese
Resume
Experience during undergraduate
Tsinghua SPPM International Case Analysis of Public Policy on SDGs
June. 2019 - Jan. 2021 / Sichuan University & Tsinghua University
As a model architect and software engineer to construct a model of interests flow among the government, hospitals, pharmacies, and third-party platforms by calculating the Shapley value in game theory, which provided a quantitative judgment indicator at the stage of helping the prescription flow model of pharmaceutical retail enterprises to take shape
Frequent Pattern Mining(FPM) Analysis
June. 2017 - May. 2018 /The NLP lab of Sichuan Changhong Company
As a research assistant to investigated the DTopkPM problem and developed an efficient algorithm, which outperformed traditional methods by spendsing less than 1000 s to discover high quality diversified top\-k patterns on a graph with 4 million nodes and 53.5 million edges.
Research on Optimal Matching of Electric Vehicl Wireless Power
Aug. 2021 - Oct. 2021 / Sichuan Univers
As a programmer and model architect develope a two port equivalent mathematical model of a series series compensated structure (PSSS) wireless energy transmission system, calculated the voltage and current transfer functions to obtain the mathematical model of the electric vehicle wireless energy system in terms of transmit frequency, matching impedance, and wireless energy transmission efficiency
Experience during graduate
Research on time series prediction with external events
Nov. 2021 - May. 2022 / Sichuan university & Shanghai Jiao Tong University.
As a core investigator( team leader) to propose a conditional causal representation (CCR) model based on causal representation learning, which solves prediction under interference has bad accuracy, robustness and generalization by extracting interference related causal representations and learning causal mechanisms subject to external interference.
Research on Class Imbalance in Credit Scoring
Oct. 2022 - Jun. 2023 / Sichuan university.
As a core investigator to propose a credit scoring model based on Temporal Convolutional Variational AutoEncoder (TCVAE) to enhance minority class samples and balance temporal dependencies.
Reaserch on LLM safety with yale and PKU
Sept. 2023 - Oct. 2023 / Georg-August-Universität Göttingen.
Contribution to the task template design, jailbreak attack and using 'GPT' as a judge.
GWGD Human brain tumour image recognition
Sept. 2023 - Oct. 2023 / Georg-August-Universität Göttingen.
As a model architect and programmer to esign a model in two days to classify graphs of human brain tumor by combination of residual and convolutional networks without pre-trained within one week, the mse of classification can be up to 78%.
Honors
Honorable
2019 American College Mathematical Modeling Competition
Third Prize
2020 Asia‑Pacific Mathematical Modeling Competition for Undergraduates
National Endeavor Fellowship(Top1%)
2019-2021 National Scholarship Assessment
Sichuan University Single Top Tier Merit-based
2018-2021 Sichuan University Scholarship Assessment
The First Prize
‘Huashu Cup’ International Mathematical Contest in Modeling
The Second Prize
4th BETT National Business English Translation Competition
The Third Prize
Mathorcup Mathematical Modeling Competition
The First Prize of Sichuan Province
2019 Contemporary Undergraduate Mathematical Contest in Modeling
Work experience
Lithium Battery Abnormal Detection
Sichuan Changhong Company:1. Anomalies labelled with tens of thousands of lithium battery images by labelme; 2. Comparison of yolov7,mixed and u2net for defect detection on lithium battery, among these methods, yolov7 get the best detection accuract, which F1 can be up to 92%
Backstage system program
Xinbo Zhichuang IOT Technology Company: 1. Developed the back-end management system of Xiaoxin Youpin and obtained the software copyright in June 2020; 2. Mainly with the help of the C++ STL library, wrote some of the functional functions, such as some memory allocation functions in the backgroundPublication
Contact
Keep In Touch
- No.24, South Section of 1st Ring Road, Chengdu
Wuhou District 610065 China - (+86) 139-8110-6646
- fengxuanzhi34@stu.scu.edu.cn
fengxuanzhi34@gmail.com