On November 14-15, the 2017 Global Home Internet Conference (GFIC) was held in Shanghai. The conference attracted 500+ companies and 3000+ professional audiences. At the GFIC2017 Global Home Internet Conference on the afternoon of the 14th, Google’s Global Product Director Dave Elliott gave an important speech on how to use media and artificial intelligence, including machine learning.
Dave Elliott first stated that TVs and computers can be seen in any media. The media can be seen in the inside of your car, including airplanes. It has spread so much. The current devices are diversified and the content will be diversified. Personal customization is very important for personal success. The extensive expansion of equipment is also a large expansion of content, and personal customization is the key to achieving uniqueness and increasing value. Now that consumer content is diversified, it can be done where and when. So in the future, companies should not only understand who their customers are, what they want to see, but also what they want to see when they want to see what they want to see.
Artificial intelligence is a very big concept. Machine learning is the simplest way to answer questions with different examples. Privately customizing some content requires machine learning and requires correct examples to answer the right questions.
Data plus your program to achieve predictions, predictions in this case refers to what the user wants to see, because Google is a software development company, has been operating for many years, but the difference is that machine learning can study the law itself . The importance of machine learning is obvious, but it is not easy to achieve. Machine learning requires a large amount of computing power. This algorithm requires a lot of computing power. For example, cloud computing, such as GPU, is also very demanding.
The example is data. Now that Google has gotten more and more data, it needs a powerful computing tool. The Kaiyuan tool is to allow you to carry out machine learning.
How can media and artificial intelligence help the public? If you want to do a private custom image, first of all not only to understand who the customer is, but also to know what is happening inside the video. Until recently, great data was still needed to understand who was doing something, but this was simplified for machine learning, which was revolutionary.
In addition, Dave Elliott explained that Google wants to create a live broadcast message to customers. If customers want to see what program, they can set different virtual information to different people. Everyone can see different TV programs and can see what they want. TV shows to watch, based on this company can find what customers really want to see.
In addition, artificial intelligence can also do some automatic highlights, such as watching a thousand hours of video, or just want to see a brief five minutes, in order to find some bright spots, you must first understand the content. A television program can be real-time, even real-time translation of subtitles, can be translated in dozens or even hundreds of languages, and finally formed a sort of content. In order to understand whether the content is appropriate, and finally to achieve the cash income and expenditure, see the advertisement inserted at a certain time, for example, see a shoe in the video, at this time there is a timely advertisement of a shoe.
Google’s latest product, TensorFlow, can help you achieve this. It was launched in November 2015. It is a Kaiyuan library and the number one development software. Compared to other Kaiyuan machine-learning products, TensorFlow is the fastest growing, it is free and can be used at any time.
Charts given by TensorFlow can simplify the whole thing. TensorFlow supports different platforms. It can even be used on ros system Android and supports different programming languages. It can make predictions, help you automatically find the highlights of the entire program, visualize all the information as a visualization to understand the customer's needs, and also understand the media itself, and then identify and understand what the customer really needs. The timing of ad placement puts the right content in the right hands at the right time.
Now in a very special era, the living room takes more and more time, tools are also increasing, Google's new products are undoubtedly refreshing, we look forward to the future Google will develop more user-friendly new products.