Here is my warm greetings! Welcome to my home page. My name is Pengcheng Xu(pronounced as Peng-cheng Hsu).
In Chinese, my name is written as 徐鹏程, where 徐 is Xu, 鹏 is Peng, 程 is Cheng. 徐 is my surname(Also has the meaning of slow). 鹏 is a giant bird in ancient Chinese myth. It has huge wings and can fly 5000 kilometers at a time. 程 means journey. So, basically my name means the journey of the huge bird 鹏.This name carries the hope of my parents for me to have a bright and shining future, as well as a promising and ambitious path ahead.
Welcome to talk to ChatGPT version of me on my PengchengGPT website 😄!
ChatGPT-Chatbot-Personal-Website repo is here.
If you think it's cool or want to use it, please star it! Thank you~
🌱 Education
🎸 Hobbies
In my free time, I enjoy singing, playing the guitar, working out, playing basketball, tennis, swimming, reading, watching films(especially sci-fi ones), and exploring everything related to science and technology. I'm particularly inspired by Richard Feynman and Elon Musk. You can check out some of my technology talks and guitar playing on My bilibili.
💻 Programming Languages
When it comes to programming languages, I'm skilled in C, C++, and Python. Additionally, I have project experience with MATLAB, Verilog, Java, R, and Javascript. I also have experience using various frameworks and libraries such as PyTorch, TensorFlow, Keras, Sk-learn, Pandas, and Horovod.
🔭 Area of Interest
My main interests are in applied machine learning for healthcare or science(bioinformatics, molecular optimization), computer vision, multi-modal learning, distributed systems and distributed learning, and computer architecture.
🔬 Research Experience
-
Molecular de-novo design through Decision Transformer and Oracle-feedback reinforcement learning
Second co-author
May 2023 - Present
Advisors: Tianfan Fu (Incoming Assistant Professor at Rensselaer Polytechnic Institute), Jimeng Sun (Professor at UIUC CS Department)
Submitted paper under review. Read the paper preprint.
-
MIRACLE: Multi-task learning based Interpretable Regulation of Autoimmune diseases through Common Latent Epigenetics
First author
Sep 2021 - Present
Advisor: Hongyi Xin, Associate Professor at University of Michigan - Shanghai Jiao Tong University Joint Institute.
Submitted paper under review. Read the paper preprint.
-
Pipe-Déjàvu: Hardware-aware Latency Predictable, Differentiable Search for Faster Config and Convergence of Distributed ML Pipeline Parallelism
First author
Feb 2023 - May 2023
Advisor: Indranil Gupta, Professor at UIUC CS Department
Read the research report
💼Work Experience
-
Software Engineer Intern
Amazon Web Service, Seattle, WA
Duration: May,2023 - Aug,2023
Worked with VMware Cloud on AWS Group.
-
Multi-modal Cognitive Computing Algorithm Intern
Shanghai AI Laboratory
Duration: May,2022 - Aug,2022
Worked on Multi-modal target detection with zero-shot depth estimation and Multi-modal Neural Architecture Search.
-
OpenMLDB - Open-source Software Developer
4Paradigm Co.,Ltd
Duration: July,2022 - Oct,2022(Part-Time)
Developed AutoFE: automated feature engineering tool. View pull request.
Presented at OpenMLDB Meetup No.7 and OpenMLDB-GitLink Code Camp 2022 mid-term presentation.
-
Deep Learning Software Engineer Intern
Intel Corporation
Duration: Nov,2021 - June,2022
Worked on Intel Neural Compressor and ML inference server software.
-
Algorithms Intern
Shukun Technology Co.,Ltd
Duration: Dec,2020 - Apr,2021
Worked on Multi-node Training for 3D-UNet with horovod.
Presented on multi-node training and horovod.
🧑🏫 Teaching Experience
-
Teaching Assistant for VE370 Intro to Computer Organization, 2021 Fall
Supervisor: Gang Zheng. University of Michigan - Shanghai Jiao Tong University Joint institute.
Recent Work
MIRACLE: Multi-task learning based Interpretable Regulation of Autoimmune diseases through Common Latent Epigenetics
(Under Review) Pengcheng Xu*, Jinpu Cai*, Yulin Gao, Ziqi Rong, Hongyi Xin. 2023. Preprint Link: arxiv.org/abs/2306.13866
Multi-task learning based interpretable gene-level methylation estimations | Research Assistant Sep 2021 - Present
Advisor: Hongyi Xin, Associate Professor of UM-SJTU Joint Institute, Shanghai Jiao Tong University
- Explored adaptable and interpretable neural network to find common genotype given 480k dimension sites, hundreds of sample.
- Designed an explainable site-gene-pathway ontology constraint to NN to discover new biomarkers by checking weights.
- Implemented a Variational Auto-Encoder to support gene-level embedding shared among datasets to obtain multi-task learning.
- Optimized a pretrain-finetune training scheme to increase accuracy by over 10%, wrote the paper under review in 2023.
Molecular de-novo design through Transformer-based Reinforcement Learning
(Under Review) TaoFeng*, Pengcheng Xu*, Tianfan Fu, Jimeng Sun. 2023. Preprint Link: http://arxiv.org/abs/2310.05365
Advisor: Tianfan Fu, Jimeng Sun
- Implemented a decision transformer architecture to improve the AUC for over fifteen molecular optimization tasks for 5% each on average.
- Applied Oracle-feedback reinforcement learning on the downstream tasks to reach higher performance than pretrained model.
- Carried out ablation study and investigation into loss curve and conditional probability over the next token as a function of previously chosen ones according to the model
Balancing Information Preservation and Computational Efficiency: L2 Normalization and Geodesic Distance in Manifold Learning
(Under Review) Ziqi Rong, Jinpu Cai, Jiahao Qiu, Pengcheng Xu, Lana Garmire, Qiuyu Lian, Hongyi Xin. 2023.
- The importance of distinguishable information in similarity measurement for unsupervised learning, manifold learning, and high-dimensional data visualization tasks.
- The limitations of conventional metrics like Euclidean distance after L1-normalization in handling high-dimensional data due to the "curse of dimensionality".
- The influence of normalization with different p-norms and the defect of Euclidean distance.
- The preservation of observation differences when normalizing data to a higher p-norm and using geodesic distance instead of Euclidean distance.
- The sufficiency of L2-normalization onto the hypersphere in preserving delicate differences in relatively high-dimensional data while maintaining computational efficiency.
- The presentation of HS-SNE, an augmentation to t-SNE based on a hypersphere representation system, which effectively addresses high-dimensional data visualization and similarity measurement intricacies.
- The better resolution of the hypersphere representation system in identifying subtle differences in high-dimensional data while balancing efficiency and computational feasibility.
Pipe-Déjàvu: Hardware-aware Latency Predictable, Differentiable Search for Faster Config and Convergence of Distributed ML Pipeline Parallelism
(To be submitted) Pengcheng Xu, Kaiyang Chen, Yuanrui Zhang, Indranil Gupta. 2023. Read the research report
Hardware-aware Latency Predictable, Differentiable Search for Faster Config and Convergence of Distributed ML Pipeline Parallelism
Advisor: Indranil Gupta, Professor of CS UIUC | Advanced Distributed Systems| Researcher | Feb 2023 – May,2023
- Implemented a predictive model that considers communication cost, model computational cost, and hardware information to predict latency and resources of parallel configurations, saving time on pre-profiling before searching the parallel configuration.
- Proposed a differentiable parallel configuration search space inspired by DARTS, can potentially reach optimal configuration faster than the original dynamic programming.
- Employed parallel random initialization using sampling algorithms like Bayesian Optimization for faster train loss convergence
Vascular Intervention Training System Based on Electromagnetic Tracking Technology
Zhikai Yang, Pengcheng Xu, Dekun Yang, Yufeng Chen, Yancong Ma. ICVRV, 2020. ieeexplore.ieee.org/document/9479727
Advisor: Lixu Gu, Professor of Biomedical Engineering, Shanghai Jiao Tong University
- Developed the framework of an augmented reality surgery training assistant system for medical student and surgery.
- Predicted the operation trajectory using LSTM and used KD-Tree to calculate the distance for operation safety warning.
- Displayed vascular model in AR with OpenGL and designed the UI interface to support translation.
- Used the aruco library in OpenCV to coordinate positioning of the QR code.
- Published Vascular Intervention Training System Based on Electromagnetic Tracking Technology on ICVRV as second author.
Get In Touch
Feel free to email me or message me.
-
Address
1010 West University Ave
Urbana, IL 61801
United States
-
Phone
217-550-1337
-
Email
px6@illinois.edu