👋 Hi there! I'm Guoying Liao, aka David. Welcome to my GitHub profile, where you'll find my adventures in coding, research, and a few personal interests.

You can reach me via email at:

• 📩 [email protected]
• 📩 [email protected]

You can reach me via a green little Chinese App:

• 👽 lgy112lgy112

💻 I have been providing code outsourcing services, so if you are interested in collaborating, feel free to contact me.

🌟 Most of my projects are completed on Lightning AI Studio. Due to client privacy requirements, I can only make a few projects public. If you need to see more, please get in touch with me.

Thank you for visiting my profile, and I hope you find something interesting here! 🙌

• Name: Guoying Liao (David)
• Hobbies: Longboard 🛹, guitar 🎸, customization 🔧, chat 💬
• Some Fun Facts:
  • 🌟 Participated in the Canton Fair twice, where I successfully traded with many overseas friends and earned my first significant income (which I spent on an iPad Pro). I have to say I hate India accent (they are good partners tho).
  • 🏥 Interned at the General Hospital of the People’s Liberation Army in Beijing, working as a physicist in the oncology department. I was responsible for dose control and QA/QC.
  • 🤓 Realized women (no offense but) slow me down coding.
  • 🧑‍🤝‍🧑 Have 10+ boy friends.
  • 🗣️ Taught 100+ on oral English for free (sometimes for a coke).
  • 🍔 Always McDonald's!

• 2020-2024: Southern Medical University, specializing in Oncology Radiology.
• 2024-: Chinese University of Hong Kong, pursuing Biomedical Engineering.

• 📄 The m6A methyltransferase METTL3 affects cell proliferation and migration by regulating YAP expression in Hirschsprung disease
• 📄 Association of rs2435357 and rs2506030 polymorphisms in RET with susceptibility to Hirschsprung disease: A systematic review and meta-analysis
  • Role: Data processing and visualization
  • Published on: 2022, as undergraduate
  • Shout out: my good fellas in Clinical College

1. Abdominal Multi-Organ Segmentation with ResUNet and ViT-UNet

   • 📂 Published Code: Lightning AI Full Development Version
   • 📂 Teaching Version on Kaggle: Kaggle Notebook
   • Key Contribution: Developed a novel loss function, DiceCELossWithKL, which improved F1/IoU scores in a very optimistic extent.
   • 🚀 Published Nova Loss on GitHub: DiceCELossWithKL

2. Ultrasound Nerve Segmentation with ViT and Superpixel Tokenization

   • 🏆 Competition: Ultrasound Nerve Segmentation | Kaggle
   • 📂 Code: vitthing (lightning.ai)
   • Innovation: Introduced superpixel tokenization for finer segmentation, utilizing image texture and color information to generate superpixel blocks of varying sizes and shapes.

3. Crested Porcupine Optimizer (CPO) in Python

   • 📄 Original MATLAB Code: Crested Porcupine Optimizer MATLAB Code
   • 🐍 My Python Implementation (proud to say I am the 1st to do): pyCPO on GitHub
   • Achievement: Migrated the CPO algorithm to Python, demonstrating its superiority in a practical project.

4. Experimenting with KAN (Kolmogorov-Arnold Network)

   • 📊 Stock Market Prediction: Stocks Prediction Tutorial, KAN_Stocks
   • 🖼️ Image Tasks (CIFAR10): KAN vs VGG on CIFAR10
   • Focus: Exploring the potential of the latest academic ideas in deep learning.

5. Many More Available on My Lightning Studio

   • 😁 You want to see you call me maybe.

6. Deep Learning Tutorials

   • 📝 Building a Language Model: NLP-Tutorial-How-to-be-Shakespeare
   • 🧠 General Deep Learning Pipeline: DeepLearning-Tutorial-Chinese
   • 🔍 Understanding KAN: Efficient-KAN-in-Chinese
   • 🖼️ ViT for Image Classification: simple-ViT-tutorial-to-recognize-taylor-and-olivia
   • 📐 Solving a Topology Math Problem: mathproblem
   • ⚡ Example of FastAi Usage: FastAi Example
   • 👽 Tutorial on MNIST Using MLP with Multi-View Training Method and KAN: MNIST Tutorial

Feel free to explore my repositories and reach out if you have any questions or collaboration ideas. Happy coding! 🎸🛹