In 2016, for me personally, the most important choices I made were in the Microsoft Research Institute (hereafter abbreviated as MSR) that has been working for 13 years, and a startup company that has been established for a few years - Face++ Technology (hereinafter referred to as Face++), the latter was selected and joined as the chief scientist. I was still living in Seattle at the time. I was really "untouched, and the news was far away." Various reports from the circle of friends both at home and abroad struck me strongly, which made me appreciate the power of the media.
Today, I have moved back to Beijing to work at Face++ for nearly six months. I am still often asked, "How is it going?", "Is Face++ the same research department as MSR?", "How is Face++ conducting research? ......" and so on. The problem is big or small, but most of them are like this. On the occasion of the New Year, I would like to share with you my observations and thoughts in the past six months, and I would like to thank you for your concern.
Next, I will focus on one of the typical issues that people are concerned about:
What are the similarities and differences between Face++ and MSR R&D?
As far as my observations and experiences are concerned, the essence of the two companies' R&D departments is almost no difference. What is the essence of a company's R&D department? I think there are three key elements that are crucial: mission orientation, staffing, and R&D. Frankly speaking, from these three points of view, I saw striking consistency on both sides, that is to say:
1) They all have the mission of positioning both product-based and cutting-edge technologies.
2) They are all gathered together with a group of elites who are pursuing extremes, have Geek spirit, and are highly self-driven;
3) They all use the same routine to advance research: identify problems --> implement, research, and understand existing methods --> conduct continuous improvement or innovation.
Among these, the most sensational scene is "human." Let me give you an example that has recently moved me: The following circle of friends that I sent on the last day of 2016 best explains the first two points of Face++'s core values: "Pursuit, Ultimate, Simple, and Reliable."
Of course, even if the two sides have the same research department, there will inevitably be differences. After all, each company has its own specific culture and management model. When the comrades around me suddenly came down from the age of thirty years old to ten years old, I felt as if I was in the middle of the elderly momentarily. (Speak a note: A colleague of mine asked me today. Why do you think the size of your mobile phone is so big? On the other hand, you feel that you are full of energy and have a heavy sense of responsibility.
What is the study of Face++'s research department?
Asked on many occasions, this issue is the most mentioned. Why do I have such questions? I want to think about two aspects of thinking. First, I want to know what areas the company specifically researches, what are the long-term topics and goals, and second, I want to understand whether a R&D department in a startup company can. To promote research in the real sense, it is still under the banner of research to do product development.
There is also an implicit misunderstanding here. After our company was widely known as Face++, more and more people mistakenly think that Face++ is just doing face technology. Face, it is indeed a business with broad prospects and continuous refurbishment. However, from the first day of its establishment, Face++ focused on three areas of application of artificial intelligence - computer vision, focusing on a series of visual identification (faces, people, objects, text, scenes, behaviors, etc.) The core algorithm to create products that can land. If you eliminate such a misunderstanding, you will understand why Face++ uses “PowerHumanwithAI†as its mission and uses “artificial intelligence technology to benefit the publicâ€. After all, the company’s full name is called defiance (English name Megvii, taken from MegaVision), which is a big vision.
Returning to the problem itself, we are currently focusing on four core issues of visual understanding (see the following figure): image classification, object detection, semantic segmentation, and sequence learning. The technical line of research is thorough deep learning: 1) Use of deep neural networks; 2) Use end-to-end learning as best as possible. Face++ should be said to be the earliest research and application of deep learning in this wave of artificial intelligence startups.
relay fuse box,12v relay box,car relay box
Dongguan Andu Electronic Co., Ltd. , https://www.idofuseholder.com