The tepid Siri did not affect the giants' love for artificial intelligence.
“I am going to replace my home Echo with Google Home,†said Rene, a product manager at a technology company, after the end of the Google I/O conference.
The same point of view, we heard no more than 5 times when Google I/O and developers chatted. They all think that the most attractive is not the Google Home smart speaker itself, but the data and artificial intelligence that Google has collected over the years has finally been combined.
For Google, the most online traffic, artificial intelligence is its basic tool.
Unlike most movie ideas, artificial intelligence does not have to have a human face or a magnetic voice. It is already something you use every day. For Google, the world's largest search engine, every search is an artificial intelligence application.
From recognizing text and speech to understanding pictures, “machine learning is not magic, it is a tool.†Greg Corrado, the founder of the Google Brain project, once said.
For each search in the search engine, the results returned to you will use artificial intelligence. It needs to first identify what the text you typed means.
This is not a simple typing keyword match, it will try to understand your problem and give a clearer conclusion. For example, when you want to know when Leonardo gets the Oscar for best actor, but he can't spell his name. Google will predict your issue. At the same time, the matching answer also joined the history of Leonardo's nomination.
Google predicts your question
When you enter the answer, you will also introduce the event background.
Now, you are seeing more and more - auto-tagging schedules in Google Mail, and even helping you to call and book hotels. You can also call Google Now to help you plan your work route. Search for "hugs" in Google Photo and you can see all the photos of the same action.
The new chat app Allo even tries to understand the social behavior of humans and automatically generates responses that you might want to use. This way you don't have to type hard, just click a button.
Google has 40,000 searches per second to generate massive amounts of data. All of this data is the basis for artificial intelligence analysis, and every search, every recommendation, every photo recognition is an opportunity to improve the algorithm.
And last year, Google also opened up its own artificial intelligence platform TensorFlow. Let more developers help improve the algorithm.
And Facebook is using the same method to challenge Google on artificial intelligence.
Recently, Facebook was exposed to manual editors to “suppress†speech and guide public opinion during the election.
But in addition, the personality recommendation algorithm presents the news and videos you are interested in in the information flow, recommending friends you may know, and these will use artificial intelligence more or less.
And Facebook has more than 1 billion people brushing social networks every day. At the same time, these people will upload large amounts of private photos and videos, many of which will not be shared publicly.
Like Google, images and videos are also the focus of Facebook.
This technology has not only stopped at distinguishing different types of animals and plants. When you search for a friend's name, it even recognizes a specific person in the video and photo.
Following the footsteps of Google's open source artificial intelligence platform, Facebook also opened up the Big Sur platform at the end of last year. Massive data processing requires smarter algorithms, and this requires experience.
Compared to Google, social networks handle the same amount of data, and it can search based on more private information.
Along the way to Siri, Amazon found a more suitable shell.
Compared with Google and Facebook, Amazon's wireless speaker Echo, which was launched worldwide in June 2015, has found a physical shape and more specific usage scenarios for artificial intelligence.
Amazon launched the wireless audio Echo
Compared to Apple's Siri, Alexa, the voice assistant behind Echo, is growing rapidly.
In addition to being able to help you call pizza, as a smart sound, Echo songs are the most convenient. Alexa can understand the music style, the name of an album, the name of a singer or combination, or even a song title.
The music library is not limited to its own music service. In addition to Amazon's own Prime Music, you can also pick up popular services like Pandora and SpoTIfy.
The day after Amazon announced the global sale of Echo in June 2015, Alexa Skills Kit, a free development tool, was also offered to developers around the world. It is hoped that third parties will explore the potential of Alexa.
Users can also participate. Echo's mobile phone application also includes a manual error correction mechanism.
When you think Alexa doesn't respond to you correctly, you can get an error. This design, which allows you to correct errors from time to time, allows users to train Alexa's intelligence to some extent. As of April this year, Echo sold 3 million units.
As a speaker, Echo is really convenient, you don't have to use your phone or enter text. Amazon chooses open source and is compatible with third parties. This is what makes Alexa an entry, not just a tool for Amazon to sell music services.
Siri has not found a scene with high frequency of use, and it is more convenient to use voice control than hand.
Siri can still do a little less, and most of them are not very practical, such as telling a joke and watching a sports result.
After being acquired by Apple, since Siri was released in October 2014, as software, it has been keeping pace with hardware updates. Every major upgrade has to wait for the Apple Global Developers Conference. It is still the rhythm of selling hardware rather than doing Internet services.
Microsoft's artificial intelligence, all can be stuffed into the body of Xiao Bing
In the Dragon Boat Festival in 2014, Microsoft's artificial intelligence robot "Xiaobing" was fired directly to the micro-envelope lock. This was the fastest-moving product after the WeChat red envelope.
Xiaobing's iteration speed is also much faster than Siri. When the third generation of Xiao Bing entered WeChat again in August 2015, she not only had access to other WeChat public accounts, but also launched a new feature every week. Such as the New Year sign, guess the name and so on.
These features seem a bit naive, but in fact they use a lot of features originally developed by Microsoft Research. For example, how-old.net, which has been passed through a circle of friends, is mainly used to identify people by judging their faces.
Future computing manipulation will be centered around users, voice assistants, and chat bots. In March of this year, Microsoft CEO Nadella concluded at the developer conference.
Xiao Bing in Weibo
Personal voice assistant Cortana The first Windows Phone to go online is now almost useless. But Xiao Bing was put on WeChat and Weibo. These are hundreds of millions of people online every month. Some people will not be able to chat with her.
Small and medium-sized companies do artificial intelligence, relying on the group
The tech circle has been talking about the fear of ultimate artificial intelligence for several years. Bill Gates, Elon Musk and Stephen Hawking have publicly stated that artificial intelligence will be the greatest threat to humanity.
In December 2015, Tesla CEO Eron Musk and YC Incubator CEO Sam Altman announced the joint research on artificial intelligence. They spent $1 billion to set up a new artificial intelligence company, OpenAI.
In an interview with well-known technology journalist Steven Levy, Musk said that if you want to prevent artificial intelligence from being used in the wrong place, the best way is to let more people use it.
In addition to worrying about the threat of artificial intelligence, OpenAI is also very valuable from a business perspective.
Technology analyst Ben Thompson once pointed out that while these people care about human well-being, the massive amount of data that OpenAI gathers will also bring direct benefits to their company. For example, upgrade some system features.
This is not difficult to understand. Tesla's AutoPilot feature requires a lot of geographic location information.
No matter whether Tesla or YC incubating companies, there is no huge amount of data like Google and Facebook.
In April of this year, Tesla CEO Maske and YC incubator led the open source artificial intelligence platform OpenAI Gym officially launched. It is equivalent to a test platform for artificial intelligence systems, such as voice assistants, image automation processing and other projects.
In the cause of rebelling against the Terminator, small and medium-sized companies are sharing data with each other and advancing their research on artificial intelligence.
Blind flanges are made of carbon steel,stainless steel and alloy steel,etc.They are used to seal or block off a pipe,like a cover or cap.Because of its good overall performance, Flange is widely used in such basic projects as chemical engineering, construction, water supply, drainage, petroleum, light and heavy industry, freezing, sanitary, plumbing, fire protection, electricity, aerospace and shipbuilding.
Type: DIN2527
Pressure: PN6 to PN64
Material: Carbon steel or other required material
Surface: Original Color, Galvanizing or Paiting as clients requirements.
Packing: Wooden Cases
Drawing and data show DIN2527 PN16, if you need other pressures, please contact with supplier.
How to order: DIN number + (size) (x d1 dimension if not ISO) + (pressure) + material if not steel
Example: for steel = DIN 2527 DN 100 PN 16 Example: for stainless = DIN 2527 DN 100 PN 16 316SS
Blind Flange,Spectacle Blind Flange,Steel Blind Flange,Stainless Steel Blind Flange
HEBEI ZIFENG NEW ENERGY TECHNOLOGY CO.,LTD. , https://www.zifengpipeline.com