The artificial intelligence technology, known for its machine learning, is now at a white-hot stage.

Ai technology, known for its machine learning, is now at a white-hot stage, as we have mentioned many times before. The technology is driving the development of computer vision, language recognition, and text analysis technologies for companies such as Google, Facebook, Microsoft and Baidu, and has become the technology base for many start-ups (some of which have been acquired before the product is released). With the development of machine learning, these successes have received a lot of media attention.

But what you're seeing is probably just a superficial phenomenon. Many studies are taking place in non-large web companies, and even the relevant research by large web companies has not been disclosed, with only a few major breakthroughs and imagenet records being paid attention to, but progress in this area has not stopped.

Just recently, Google's DeepMind team reported on their early work in building a so-called "neural network Turing" algorithm to create a system, and Facebook has shown a "generic" video analytics 3D feature that Microsoft researchers believe may be a breakthrough for some types of depth learning algorithms.

We're going to discuss some of the results in next year's conference structure data conference, but who wants to really understand the progress of the deep learning area, the Nips meeting in Montreal, Quebec, may give you an answer last week. This year-by-year conference is increasingly occupied by deep learning. Of the 411 papers received this year, 46 of the top 100 commonly used words contain the word "deep" (according to the theme model of Stanford University doctoral student Andrej Karpathyde), which is twice times 23 last year.

At a separate depth study seminar, the number of posters soared from 28 last year to 47 this year. Although some of the larger research breakthroughs have been reported (e.g., two types of neural network combinations of karpathy studies automatically generate image titles), other potential important work has not aroused widespread public concern.

Yoshua bengio--, a researcher at the University of Montreal in Canada who is well known in the deep circle of learning, has so far refused to charm the company's laboratories-his team seems to be busy. Bengio is listed as one of the 5 co-authors of this year's Nips paper, but his name does not often appear in the news about Skype's translate or Facebook's attempt to check out the drunk photos released by users.

In a recent TEDx conversation, Enlitic CEO Jeremy Howard talked about the advances in translation and medical imaging, and he also showed how the software developed by the company helped doctors train computers to classify medical images in just a few minutes.

What I want to emphasize here is that I am not saying how much research is being done on machine learning at the moment, nor is it warning that humans are being taken over by AI. I just want to remind you that in fact, many of them are underestimated by the media, but still, people in this field should keep up with the times.

If our infrastructure is preparing for the impact of machine intelligence, then legislators, national security agencies, ethicists and economists need to be aware of what is happening and what will happen, no matter how it is defined. (In another field of AI research, Allen is pouring money into a project that tries to give the computer practical knowledge)

Text Link: What We read about deep learning is ethically the tip of the icebergs (Zebian/Wei)

Free Subscription "CSDN cloud computing" micro-letter public number, real-time grasp first-hand cloud news!

As the most professional cloud computing service platform in the country, CSDN provides cloud computing, large data, virtualization, data center, OpenStack, Cloudstack, Hadoop, Spark, machine learning, intelligent algorithms and other related cloud computing views, cloud computing technology, cloud computing platform, cloud computing practice, Cloud computing industry Information Services.