I am Seunghyeop,
an AI researcher
Bachelor
in Konkuk University.
About
I am an undergraduate student at Konkuk university
Currently, I am focusing on being a researcher in Machine Learning and Artificial Intelligence as well as developing applications.
I have a lot of interest in cryptocurrency.
I am a bright and outgoing person. I consider being a nice and polite person is most important. My listening attitude and arbitrating conflicts made me who I am now.
Expertise
- Multi-Agent Reinforcement Learning
- UGV & UAV navigation control
- Deep Learning
- Image Classification
- Cryptocurrency trading with RL
Experience
DMS Lab
Undergraduate Researcher
JUNE 2021 - Present
I am currently studying and working at DMS-Lab(Distributed Multimedia Laboratory). The researches that I'm working on are UGV & UAV control with Multi-Agent Reinforcement Learning and classifying disease of agricultures using Deep Neural Network (Deep Learning) Image Classification.
DMS-Lab WebsiteOXEN
Konkuk University Groupsound
March 2021 - PRESENT
What I consider as important as studying is a hobby which is music for me. I love to listen to various genres of music such as metal, blues, jazz, funk. I also love to play an instrument.
So I play bass guitar in a band called OXEN.
Bandleader for 2021-2022.
VEX Robotics
Robot Programming
January 2016 - April 2017
VEX Robotics Competition World Championship is one of the biggest robotics competitions in the world. My role in the team was designing robots, programming robots and planning the strategy of the game.
Even though I didn't win a high rank, this experience brought me inspiration to be a researcher of computer science and robotics.
Education
Konkuk University
Bachelor in Computer Science Engineering
MARCH 2019 - PRESENT
Major in computer science engineering at Konkuk University.
Computer science engineering(hereinafter referred to as CS) and IT development were attractive to me since I was young. Rather than other majors, implementation in CS is quite instant. This means, whatever I think, I can implement applications or apply new technology and method as soon as I am inspired.
So I'm focusing on improving my coding skills and way to think in a logical way as well as learning research skills that can improve my CS career.
Se-Kwang High School
Science-Focused High School
MARCH 2015 - MARCH 2018
I completed the regular curriculum of Korean high school without missing a single day for 3 years.
Took scholastic aptitude test(for university admission) at NOV 2017 and 2018.
Coding Language Skills
Recent Works
Here are some of my favourite projects and experiences I have done lately. Feel free to check them out.
Disease detection for Rapa
In order to prevent the spread, respond quickly, and reduce pesticides in the disease of cabbage, Image Classification is improved through Transfer-Learning and modifying deep neural networks layers to increase accuracy.
This can be useful especially in Korea because Korea consumes lots of cabbages than other countries.
- Deep Learning
- Agriculture
Mirrormoji
This idea of the project was inspired by the research which shows usual emotion affects a lot from dressing style.
Attach mirror film on display and display shows time, news headline, schedule and weather forecast. And when the user calls "Mirrormoji" like "Hey Siri!", the user can start a conversation with Mirrormoji and ask for style recommendations.
If users ask for style recommendations, Mirrormoji recommends style based on users current emotion with face, voice and words that users use with ER-Engine(Emotion detection AI) supported by LG ThinQ.
- Product
- Development
- LG ThinQ
Stock market trading with RL
With reinforcement learning, the agent learns about KOSPI(Korea Stock Price Index) with the economic indicator. An agent can do 3 actions which are buying, selling and holding stocks and after the training, the agent will predict the stock market and do actions to take profit.
- Research
- Reinforcement Learning
Covid diagnose through Chest X-Ray
I have experience in creating the best neural network for chest x-ray datasets. I built a distributed processing environment with 7 servers through MATLAB, succeeded in learning twice as fast, and learned how to solve high-performance computing power essential for AI through distributed processing.
- Research
- Deep Learning
- Distributed System Computing
Self Driving mobile robot through DRL
The performance of self naviation was poor by using purely DRL. mixture of SLAM and Path Planning algorithms with DRL solve this problem. The architecture was divided into three modules to create a map with SLAM, then ACO algorithm created a path to destination and a robot that flexibly avoids dynamic obstacles through the DRL.
- Graduation Project
- Research
- Reinforcement Learning