Profile
- Research Engineer at OpenAI
- CV : [PDF] (as of Mar. 2019)
Links
Email: | carpedm20 at gmail.com |
---|---|
Github: | github.com/carpedm20 |
Twitter: | twitter.com/carpedm20 |
Linkedin: | linkedin.com/in/carpedm20 |
Slideshare: | slideshare.net/carpedm20 |
Google Scholar: | author page |
Research Experience
OpenAI (2018.09 ~ )
- Reinforcement Learning / Robotics / Language (Advised by John Schulman)
University of Southern California (2016.12 ~ 2018.09)
- Reinforcement Learning (Advisor: Prof. Joseph J. Lim)
Seoul National University (2018.04 ~ 2018.07)
- Computer Vision (Advisor: Prof. Bohyung Han)
Statistical Artificial Intelligence Lab @ UNIST (2014.09 ~ 2015.10)
- Natural Language Processing (Advisor: Prof. Jaesik Choi)
Lawrence Berkeley National Laboratory (2015.07 ~ 2015.09)
Work Experience
Devsisters, Research Engineer (2016.04 ~ 2018.08)
Vingle, Software Engineer (2015.10 ~ 2016.04)
Moloco, Software Intern, Palo Alto, USA (2014.10 ~ 2015.01)
Awards
Machine Learning
- AI Grant Fellowship ($2,500 cash + $20,000 credit as award) :: 2018
- DataCom 2015 Best Paper Award :: [PDF]
Supercomputing
- International Supercomputing Conference 14 Student Cluster Challenge :: Finalist
- Asia student Supercomputing Challenge 14 :: Finalist
- Asia student Supercomputing Challenge 13 :: Finalist
Hacking
- 2013 Korea Whitehat Contest :: 3rd place (reward $8,000)
- Awarded by the Minister of National Defense (국방부 장관상)
- Holyshield 2013 :: 1st place (reward $1,000)
- 2014 Korea Whitehat Contest :: Finalist
Scholarships
- Global Scholarship for Undergraduate Research Opportunities Program, UNIST :: 2015
- National Science and Engineering Scholarship, Korean Student Aid Foundation :: 2013
- Academic Performance Scholarship, UNIST :: 2011-2015
ETC
- 2016 The Best Undergraduate Award (미래창조과학부장관상)
- Ranked 1st out of 509 undergraduates, awarded by the Minister of Science and Future Planning
- 2014 Student Outstanding Contribution Award, awarded by the President of UNIST
- 2013 Student Outstanding Contribution Award, awarded by the President of UNIST
Personal Projects
Machine Learning
Generative Models
- DCGAN-tensorflow :: Deep Convolutional GAN :: code :: demo
- SPIRAL-tensorflow :: Synthesizing Programs for Images using Reinforced Adversarial Learning :: code
- simulated-unsupervised-tensorflow :: Simulated and Unsupervised learning with GAN :: code
- DiscoGAN-pytorch :: Discover Cross-Domain Relations with GAN:: code
- BEGAN-tensorflow :: Boundary Equilibrium GAN :: code
- BEGAN-pytorch :: Boundary Equilibrium GAN :: code
- pixel-rnn-tensorflow :: Pixel Recurrent Neural Networks :: code
- visual-analogy-tensorflow :: Deep Visual Analogy-Making :: code
Reinforcement Learning
- ENAS-pytorch :: Efficient Neural Architecture Search via Parameters Sharing :: code
- DQN-tensorflow :: Human-Level Control through Deep Reinforcement Learning :: code
- deep-rl-tensorflow :: 1) Prioritized 2) Deuling 3) Double 4) DQN :: code
- NAF-tensorflow :: Continuous Deep q-Learning with Model-based Acceleration :: code
- a3c-tensorflow :: Asynchronous Methods for Deep Reinforcement Learning :: code
- text-based-game-rl-tensorflow :: Language Understanding for Text-based Games using Deep Reinforcement Learning :: code
Program Induction
- program-synthesis-rl-tensorflow :: Leveraging Grammar and Reinforcement Learning for Neural Program Synthesis :: code
- NTM-tensorflow :: Neural Turing Machine :: code
- pointer-network-tensorflow :: Pointer Network :: code
- neural-combinatorial-rl :: Neural Combinatorial Optimization with RL :: code
NLP
- MemN2N-tensorflow :: End-To-End Memory Network :: code
- variational-text-tensorflow :: Neural Variational Inference for Text Processing :: code
- lstm-char-cnn-tensorflow :: Character-Aware Neural Language Models :: code
- attentive-reader-tensorflow :: Teaching Machines to Read and Comprehend :: code
Speech Synthesis
- Multi-Speaker-Tacotron-Tensorflow :: Deep Voice 2: Multi-Speaker Neural Text-to-Speech :: code :: demo
- Neural Faces :: Human faces from Neural Network :: code :: demo
- Poet Neural :: Poet born from Neural Network :: code :: demo
- Fake Review Generator :: Fake review generator born from Neural Network :: code :: demo
- ReviewDuk :: Korean Sentiment Analyzer in Morphologically Rich Languages :: code :: demo
- FoodClassifier :: Food image classifier :: python :: matlab :: report
- word2vec-tensorflow :: Word2Vec implementation :: code
Hacked services
- LINE :: Mobile Application Reverse Engineering :: PoC :: PPT
- KakaoTalk :: Mobile Application Reverse Engineering :: PoC :: PPT
- Between :: Windows Application Reverse Engineering :: PoC
- Ndrive :: Web Protocol Reverse Engineering :: PoC
- Korail :: Mobile Application Reverse Engineering :: PoC
- UNIST :: Reverse Engineering & Network Exploit :: Report
- yes24 :: Secret :: Article
Visualization
Open Source Contribution
2015
- PaperTrend :: Analyzing Trends in Research Papers :: Flask & React.js
- awesome-hacking :: A curated list of awesome Hacking tutorials, tools and resources :: Hacking
- Chrome Email Checker :: Chrome extension:: Chrome web store
2014
- MovieTag :: Django :: Machine Learning
- emoji :: Python library
- UNIST Bus When? :: Chrome extension :: Chrome web store
2013
- MovieDuk :: Django
- HeXA Bot :: KakaoTalk Messenger bot
- UNIST Robot :: Windows program (C#)
Facebook bot
- Let's Work CS :: 2,954 users :: Facebook bot :: Facebook link
- Let's Go Abroad CS :: 2,886 users :: Facebook bot :: Facebook link
- 15 minutes Before Lunch :: 1,541 users (1 of 3 UNIST students used) :: Faecebook bot :: Facebook link
- UNIST Portal Bot :: 1,381 users (1 of 3 UNIST students used) :: Facebook bot :: Facebook link
- UNIST FedEx :: 785 users (1 of 4 UNIST students used) :: Facebook bot :: Facebook link
Presentations
2018
- 머신러닝 해외 취업 준비: 닳고 닳은 이력서와 고통스러웠던 면접을 돌아보며 (SNU 2018, 18-07-18)
- Random Thoughts on Paper Implementations (KAIST 2018, 18-04-25)
2017
- 책 읽어주는 딥러닝: 배우 유인나가 해리포터를 읽어준다면 (DEVIEW 2017, 17-10-16) [Video (Korean)]
- 알아두면 쓸데있는 신기한 강화학습 (NAVER Seminar 2017, 17-08-16) [Video (Korean)]
2016
- 딥러닝과 강화 학습으로 나보다 잘하는 쿠키런 AI 구현하기 (DEVIEW 2016, 16-10-25) [Video (Korean)]
- 지적 대화를 위한 깊고 넓은 딥러닝 (PyCon APAC 2016, 16-08-14) [Video (Korean)]
- 텐서플로우 설치도 했고 튜토리얼도 봤고 기초 예제도 짜봤다면 (TensorFlow Korea meetup, 16-06-15)
Old
- 140분의 파이썬 (NAVER D2 대학생 세미나, 14-08-26)
- 쉽게 쓰여진 Django (14-03-19)
- 영화 서비스에 대한 생각 (14-03-05)
- 카카오톡으로 여친 만들기 (13-06-29)
Skills
- Programming: Python, Lua, Go, node.js, Ruby, Java, C++, C#
- Human Languages: English - Professional working proficiency, Korean - Native