GAMMALAB at BUPT

• Participated in the development and contribution of OpenHGNN, an open source graph neural network library, and has received 746 stars on GitHub.
• During this period, I independently completed an in-depth analysis of hierarchical reasoning heterogeneous graph neural networks for fact verification, and successfully reproduced and integrated the previous ICML2020 research on efficient probabilistic logical reasoning based on graph neural networks.
• Received a recommendation letter from Professor Shichuan.

Data visualization sentiment analysis system Emoprobe

GitHub: Emoprobe，demo：Emoprobe

• Reproduce the TESAN model of cutting-edge papers, deploy it to the production process through parameter adjustment and optimization, and provide corresponding APIs for calling to achieve compatibility with multiple systems and different computing devices.
• We used crawling to create a self-made data set and called the Chinese word vector model to create a Chinese comment data set that is highly suitable for the task.
• In a production environment, one-of-six social media sentiment classifications achieved an accuracy of over 70%.
• Optimize back-end requests so that the server can efficiently complete emotional reasoning for massive reviews with less memory overhead without the need for a GPU.
• Through teamwork, a highly modular visual sentiment analysis system deployed on the server was realized.

A kind of deep Q network game dice game design based on incomplete information game

• Based on the ideas of asymmetric imperfect game theory and Q-learning in reinforcement learning and deep Q network technology, the corresponding game AI is able to compete with human players at a high level.
• As the project leader and main developer, I am responsible for organizing the division of labor and project promotion, and ultimately enables the project to support visualization and local area network connection.
• GitHub: LiquorDiceGame

Machine Learning

• Proficient in the basic theories of machine learning and able to practice common algorithms, such as linear regression, support vector machines, decision trees, Chessian models, and neural networks.
• Have rich practical experience in complete workflows of machine learning and data mining.
• Be able to skillfully use libraries such as pandas and numpy for data preprocessing and feature engineering, and be able to efficiently use frameworks such as PyTorch and TensorFlow to build and tune mainstream deep learning models, and be able to use cutting-edge papers based on Theory, reproduce its model and code, and be able to modify it according to reality and put it into the production environment.
• Have practical experience in parameter adjustment of deep learning models such as Deepfake and Stable Diffusion.
• Master basic theoretical knowledge in the fields of graph neural networks, knowledge graphs, computer vision, and natural language processing, and have practical experience in related projects.

Electronic information engineering

• Ability to design and debug analog/digital circuits
• Have practical experience in digital signal processing/digital image processing projects
• Master verilog language writing
• Able to perform assembly language programming and simple microcontroller development
• Proficient in Python and C++ programming languages and object-oriented programming
• Familiar with Git and Docker, and good at code management, teamwork and software deployment.

Server operation and maintenance

• NCRE Computer Network Technology Level 3, master LAN network configuration
• Set up a home server by yourself and perform operation and maintenance, and have professional capabilities in server management, fault diagnosis, performance tuning and LAN network configuration
• Familiar with basic Linux operations
• Proficient in virtualization technologies such as Docker and virtual machines
• Master basic front-end and back-end debugging technology and MySQL database technology

Teamwork and communication

• Have excellent English communication and literature reading skills, be able to communicate fluently across cultures, and collaborate effectively in a multicultural environment.
• Have rich practical experience in leading small development teams to promote projects.
• Leader of the Lianchuang team, he has organized a team of more than 20 people for long-term technology sharing and game project development.
• CET-4: 634
• CET-6: 569