About me
Welcome to my personal website! I am Keru Chen, currently a junior majoring in Automation at Xi’an Jiaotong University. I am set to receive my Bachelor of Engineering degree in the fall semester of 2025.
My research interests span the field of artificial intelligence, including the theory and application of machine learning, particularly in reinforcement learning, robotics and LLMs. I have also worked on projects in computer vision and computational optical imaging.
I am very fortunate to be advised by Prof. Sen Lin from Computer Science Department at University of Houston. I am currently a research intern at Milab, Westlake University, specializing in robotics, reinforcement learning, and large language-vision-action models.
You can find my CV here: Keru Chen’s Curriculum Vitae.
Email / Github / Zhihu / 500px / BiliBili
Education
Bachelor of Engineering in Automation (Sep 2021 – Present)
- Xi’an Jiaotong University
- GPA: 3.4/4.3
- Major courses: Data structures, Programming design, Operational research, Artificial Intelligence, University Physics, Pattern recognition, Computer graphic, Reinforcement learning, Digital Signal Processing, Principle of Automatic Control, Probability Theory and Stochastic Process
Research Experience
Westlake University (Apr 2024 – Present)
- Supervised by Prof. Donglin Wang from Westlake University.
- Design noval algorithm based on vision-language-action model(e.g. OpenVLA) using RL.
University of Houston (Sep 2023 – Present)
- Supervised by Prof. Sen Lin from University of Houston.
- Combining jump start reinforcement learning and constrained reinforcement learning, the algorithm is designed and analyzed theoretically.
University of North Carolina at Chapel Hill (Jan 2024 – May 2024)
- Supervised by Prof. Tianlong Chen from Broad Institute of MIT and Harvard, who will soon join The University of North Carolina at Chapel Hill as an Assistant Professor of Computer Science.
- Take part in research about time series forecasting and lead research on attacking LLM&RAG. You can find our paper here.
Project
Translation of Breast Cancer Cell Staining Images Based on Multi-Scale GAN
- Supervised by Prof. Yan Yang from School of Mathematics and Statistics, Xi’an Jiaotong University.
- Received the National Second Prize at the 9th National Undergraduate Biomedical Engineering Innovation Design Competition.
- We used a U-Net with an added attention mechanism as the generator, a multi-scale CNN as the discriminator, and a loss function based on the Frobenius norm to balance the weights of each discriminator, all within the training framework of a Wasserstein GAN.
Waveformer: A Transformer based EEG Sleep Stage Classifier
- Supervised by Prof. Gang Wang from The Biomedical-Information Engineering laboratory of State Ministry of Education, School of Life Science and Technology, Xi’an Jiaotong University.
- Combined with deep learning and signal processing techniques, i.e. encoder block of Transformer, convolutional auto-encoder and wavelet transform to classify sleep stages using EEG, and reached a classification accuracy of 75% in a private dataset, the highest level in this course.
- This project and its report have been uploaded to Github.
Face super-resolution based on deep learning
- This project is participating in National Innovation and Entrepreneurship Competition.
- Supervised by Prof. Jingang Shi from School of Software Engineering, Xi ‘an Jiaotong University.
- Improved IDPT(IJCAI 2022) algorithm and build web-demo based on Flask and Ajax according to the algorithm to realize face detection, crop and super-resolution restoration of a given photo.
Get in Touch
If you want to get in touch, you can reach me at any time via chenkeru0115@gmail.com or chenkeru@stu.xjtu.edu.cn .