This is Ronnie Wang, a software development engineer specialized in data science and machine learning. I graduated as a Master from South China University of Technology and Polytech de l’Université de Nantes.

My research field during my master degree is online handwriting recognition. However, as a general algorithm engineer, I touch more than one machine learning field, such as language understanding, search ads CTR prediction, recommendation system, computer vision and so on.

The fastest method to build an impression on me is by regarding me as a typical Otaku+geek, as I have been telling other people that I watch animes. However, like the first project vector in PCA, this can fast build a general impression, but not accurate. In fact, I rarely watch animes in recent years, and I never read comics or light novels. I’m more interested in anthropology, because the evolution of human culture fascinated me. Besides, to someone’s surprise, I’m very conservative about the evolution of information technology. I never believe in AI, in which field highlights producing toys to pleasure human beings. Electronic devices are bringing a much more ridiculous lifestyle than ever before, and you can have a glance at it through Black Mirror. Infrastructures weights more than surface technologies such as auto cars. I admire those geeks who has solid background in infrastructure development, especially those full-stack developers.

And, especially, I dislike Deep Learning very much. Thanks to my blind mentor at SCUT who knows nothing but keeps pursuing high recognition rate in handwritten characters, I learned nothing from him and his neural networks. He showed no respect to students who are dedicated to learn something intuitive, and I suppose that no one in his lab will know anything other than deep learning soon. Only by learning traditional machine learning techniques can one understand how this fake magic actually works! Actually, my boldest dream is to slap him on his face on a top academic meeting with traditional machine learning.


留言

2016-11-01