BIGBALLON
Me
Wei Li
M.S. Student [at] NCTU
fm.bigballon [at] gmail.com
Just for fun.
I am interested in Deep Learning,
(Deep) Reinforcement Learning
as well as Computer Game AI .
I aim to combine the great power of
Games and Deep Learning.

Biography

Wei Li (李韡) is a Computer Science M.S. student in Institute of Computer Science and Engineering at National Chiao Tung University(NCTU), advised by Prof. I-Chen Wu. He is a member of Computer Games and Intelligence(CGI) Lab . His research interests lie in the field of Deep Learning, (Deep) Reinforcement Learning and Computer Games AI. He has a passion for sharing his learning experience with others.
Feel free to contact him if you have any questions!

Timeline

2022 Mar.
2 paper accepted to CVPR'22
U2PL: Semi-Supervised Semantic Segmentation Using Unreliable Pseudo Labels
UniVIP: A Unified Framework for Self-Supervised Visual Pre-training
2021 Dec.
Raised another cat
We got a second cat. (321)
2021 Seq.
1 paper accepted to NeurIPS'21
MST: Masked Self-Supervised Transformer for Visual Representation
2020 May.
Raised a cat
A Chi's (ice cream) broke into my life
2019 Aug. - Now.
Researcher @ SenseTime
Smart City Group
2019 - Outstanding award
Material: [distribuuuu] [vscode]
2019 Mar. - 2019 Aug.
Research Intern @ SenseTime
Face Anti-spoofing
2018 Feb. - 2019 Mar.
CZF: CGI Zero Framework
Lightweight, Scalable AGZ/AZ Training Framework
Easy to develop new games based on AGZ/AZ
Allowed to be extended to a more general framework
WahahaNoGo is the strongest NoGo Program, which
got the TAAI 2018 NoGo Champion
got the ICGA 2019 NoGo Champion
Proposed a new training method called MPV-MCTS
Paper (Multiple Policy Value Monte Carlo Tree Search) was accepted to IJCAI 2019
Material: [slide] [paper] [AGZ paper] [RL paper] [caffe2]
Supervisor: Prof. I-Chen Wu
2018 Mar. - 2018 Jun.
A Survey of Learning Rate Schedule
Proposed a scheduling method on classification
Paper (Stochastic Gradient Descent with Hyperbolic-Tangent Decay on Classification) was accepted to IEEE WACV 2019
Material: [code] [paper]
Supervisor: Prof. I-Chen Wu
2017 Seq. - 2018 Feb.
Teaching Assistant @ MediaTek Inc
2017 Fall - Deep Learning Training Program
Instructor: Prof. I-Chen Wu
2017 Jun. - 2017 Sep.
Research Volunteer @ Project-Theta
Defect Inspection & Automated Optical Inspection
Open source: [cifar-10-cnn] [CIFAR-ZOO]
Supervisor: Prof. H.T. Kung, Academia Sinica
2016 Oct. - 2017 July.
Development of Chinese Chess AI
CGI Lab Chinese Chess Group
Won the 2nd place in TAAI 2016 Tournaments
Won the 1st place in ICGA 2017 Tournaments
Supervisor: Prof. I-Chen Wu
2016 Sep. - 2019 Jun.
Graduate student @ NCTU
Computer Games and Intelligence(CGI) Lab
Game AI, Deep Learning, Reinforcement Learning
Supervisor: Prof. I-Chen Wu

More

↑