Chi-Chun Chen - Personal Portfolio

About me

🏫
I am a Master's student studying Statistics at University of California Davis (UCD) starting in September 2024.
I graduated with a B.S. in Statistics and Information Science, with a minor in Computer Science and Information Engineering, from Fu Jen Catholic University (FJCU), Taiwan, in February 2024.
♥️
My interests lie in Data Analysis, AI algorithms, Machine Learning, particularly in Computer Vision and Large Language Models (LLM).
I also have experience as a data science intern. I welcome discussions on related research topics! 🏢
Previously, I worked as an Assistant AI Alogorithm Engineer specializing in AI algorithm development at the Industrial Technology Research Institute (ITRI) in Hsinchu, Taiwan.

🎉🎉🎉 Updates & Achievements!!!

ICIP Paper accepted!!!

Our paper has been accepted by ICIP 2024, see you in Abu Dhabi!
📥 Download My App.

New Chat Mode! Healing Chat can help heal your mind through anonymous chatting!!!!!
Selected Works

Time prediction analysis and discussion Competition of Taiwan national highway smart traffic management creative competition
📥 Download My App.

New Chat Mode! Healing Chat can help heal your mind through anonymous chatting!!!!!

📥 Download My App.

New Chat Mode! Healing Chat can help heal your mind through anonymous chatting!!!!!

Resume

Education

University of California, Davis Expected: Sep.2024 - Jun.2026
• M.S. Statistics
Fu Jen Catholic University
Sep.2019 - Feb.2024
• Major: Statistics and Information Science
• Minor: Information Science and Computer Science
• Elective: Big Data Industry and Education Project

Work Experience

AI Algorithm Assistant Engineer Mar.2024 - Aug.2024
Industrial Technology Research Institute See more details
Data Scientist Intern May.2023 - Sep.2023
Industrial Technology Research Institute See more details

Deatils

AI Mentor

The AI Mentor system integrates domain-specific knowledge and the experience of top salespeople and good instructors when interacting with clients. By building this AI Mentor system using ChatGPT, the goal is to train new staff through the AI Mentor, shortening the time-consuming process of hands-on teaching. There are two function of the AI Mentor system, including:

● Instructional Guidance: Insurance sales trainees can ask questions through the instructional guide, and the simulated human-like AI instructor will provide the correct product information or relevant knowledge in response.

● Simulated Practice: The AI simulates real-life customers, offering a genuine training experience. After the practice, the system provides scores and suggestions for improvement, allowing insurance sales traines to accumulate conversation experience and enhance their practical skills.

Contributes:
1. Research about Open-Domain Question Answer papers, provide better innovation suggestions to DRQA system.
2. Participate in the development process of the system

Mastu CCTV

Analyze the monitor image in the vistor centers in Mastu, predict pedestrian flow in order to fine-tune the monitor image recognition system model.

Beigan Visitor Center

Nangan Visitor Center

Dongyin Visitor Center

Juguang Visitor Center

Business data analysis, testing and report

The AI model predicts future sales and inventory levels of products.
Use traditional Moving Average methods and Deep Learning Models to predict future sales of products.

Compitions and Papers

Paper accepted by ICIP 2024 Jun.2024
See more details
International ICT Innovative Services Awards 2022 Nov.2022
Team Leader
👑 The First Place of Information Innovation team 8.
👑 The Best Popularity Award

Click to preview Firdt Place Award!

Click to preview Best Popularity Award!

Click to preview

Click to preview

PDF Preview

Download

PDF Preview

Download

PDF Preview

Download

Video Preview
International ICT Innovative Services Awards 2022
In a fast-paced society, it isn't easy to find good friends, so this research aims to create a chatbot exclusive to the user. Unlike common chatbots or dating apps on the market, our research's chatbot, the Language Style of Chatbot, reply's sentences will be converted through the "Language Style Model". This model is innovatively created by this research. The language style of the chatbot will be conformed to the user's preference, and the reply will instantly break the time limit. The "Language Style Model" combining N-gram, Sentence Deconstruction model, and Word2FunctionVector. The sentence deconstruction model and Word2FunctionVector were innovatively established for this research. Its functions are briefly described below:

1. Sentence deconstruction model: Its function is to learn the language style of an author. Its operation process first is to segment the author's article into sentences and words, then convert the terms according to the part of speech, and finally record them in the model.

2. Word2FunctionVector: Similar to the usage of Word2Vec, it can convert a set of words into a set of vector spaces to express their semantic meaning, and the vector space given to terms by Word2FunctionVector can better represent their usage. For example, the words "I" and "you" have different meanings, and their parts of speech are personal pronouns, and their use is very similar. For example, "I am delighted" and "You are pleased" both express that someone is happy, but the words of these two sentences are not the same. In this case, the vectors assigned to the two words by Word2FunctionVector will be similar to the vector given by Word2Vec.

Also, our research can use the MBTI test and three-stage multiple-choice questions to customize a user-specific chatbot. The options in the three-stage multiple-choice questions are all sentences transformed by the Language Style Model. The Language Style Model improves the problem of standardized responses of chatbots on the market and better meets user preferences to create their chatbots.
See more details