Yanghua Jin is a deep learning engineer at Preferred Networks, Tokyo, Japan. Before that, he received his bachelar’s degree from Fudan University, Shanghai, China. His research interest focuses on applying generative models to Japanese anime.

Email: jinyh [at] preferred [dot] jp

Timeline

Crypko.AI, Founder
2018.5 - 2018.10 Building next-generation cryptocollectible game with GAN and Ethereum.
Director, Machine Learning Development, Backend Development Currently, this project belongs to Preferred Networks.

Fudan University, B.S. with honors in Computer Science, Shanghai, China
2013 - 2018
Student of honors program.

Kyoto University, General Exchange Program, Kyoto, Japan
2016.9 - 2017.9
Adviced by Prof. Kashima Hisashi.

Preferred Networks, Tokyo, Japan
2017.3 - 2017.8
Researched on line drawing colorization and image style transfer models.

Morgan Stanley, Technology Summer Analyst Program, Shanghai, China
2016.7 - 2016.9
Designed and implemented an Online Latent Dirichlet Allocation based email topic analysis system for Mircosoft Outlook, and built several supervised and unsupervised applications based on learnt topic distributions.

Invited Talk

Exploring the Anime Characters Creation with Generative Adversarial Networks
Deep Learning: Theory, Algorithms, and Applications, Tokyo, Japan, Mar 19-22, 2018 Video is available on YouTube

Publications / Manuscripts

Towards the Automatic Anime Characters Creation with Generative Adversarial Networks
NIPS 2017 Workshop on Machine Learning for Creativity and Design.
Yanghua Jin, Jiakai Zhang, Minjun Li, Yingtao Tian, Huachun Zhu
[paper][web site]