Here is Weijie Shi (Jason Haggard).

I am a second year postgraduate in Computer Science and Technology at Soochow University (Suzhou, China) with GPA 3.8/4.0. I am extremely fortunate to be advised by Jiajie Xu.

Research Interests

Key words: spatial and temporal data, AIGC, knowledge graph, Multimodal

My previous research focuses on practical problems that artificial intelligence faces in real life. My interests are on the Machine Learning and its applications in Traffic System (such as map-matching, trajectory similarity computation). In a word, advanced technologies like ML and Traffic System positively influence the life of everybody (e.g. security tracking, traffic management, etc.).

Recently, with the booming development of large language models (LLMs), we have been amazed to witness the barriers between machines and natural languages being broken down. Many fascinating projects have utilized LLMs as controllers to create powerful automated and integrated tools (such as HuggingGPT, Visual ChatGPT, AutoGPT, ToolFormer). The combination of various tools with LLMs has led to a further burst of vitality. However, the fatal flaw of LLMs is their inability to guarantee trustworthy responses. Extracting knowledge from the internet, storing knowledge in a structured manner, and indexing knowledge to assist large models will greatly enhance the LLMs’ ability to understand and express knowledge. I wish to devote my talent to this meaningful cause and bring well-being to society.

Personal Skills

My engineering skills are well-honed and complemented by a strong intuition for scientific research. Moreover, I excel in feature engineering techniques. I can also handle the big project and contribute to significant improvements in model accuracy by combining specific questions.

1. Proficient of mainstream deep learning methods: Transformer, Diffusion, RL, GAN, VAE, CNN, RNN, and GCN.
2. Experiential with integrated learning frameworks: XGboost, Catboost, and LightGBM.
3. Proficient of languages: Python, Pytorch.
4. Slightly understanding of tools: Pytorch Lightning, Scala, Spark, Docker.
5. Used tools: Docker, Neo4j, C++, Java, Tensorflow, Keras, C#, Vue, Tornado/Flask, Bash.

Academic Publications

• ICDE 2023 - The 39th IEEE International Conference on Data Engineering

  “LHMM: A Learning Enhanced HMM Model for Cellular Trajectory Map Matching”

  Weijie Shi, Jiajie Xu, Junhua Fang, Pingfu Chao, An Liu, and Xiaofang Zhou

  The description of this paper in Here

  Download PDF

  Access Source Code

• 专利、软著

  一种面向信令数据的定位与路网匹配方法及系统 专利号：2023103350720

  基于seq2seq模型的智能聊天机器人系统 登记号：2020SR1702072 Certification

Projects

You could clink Projects to turn to the page of Projects.

Awards

• Tianchi Competition, 5nd place out of 670 (CAAI-BDSC2023 Social Graph Link Prediction Task 1: Small Sample Scenario Link Prediction), May 2023 Competition Link

• Tianchi Competition, 139th out of 3493 (Imagine Computing Innovation Technology Competition Track 1: Edge Cloud Content Distribution Network Customer Experience Prediction Algorithm), November 2022 Competition Link Certification

• Lanqiao Cup, Jiangsu Province First Prize, August 2020

• National Computer Design Competition, Second Prize, May 2020 Certification

• American College Student Mathematical Modeling Competition, Meritorious Winner, March 2020 Certification