Artificial intelligence generally appears as a tool to assist human work. Sweeping robots, medical robots and waiter robots are the most common forms of artificial intelligence. In fact, artificial intelligence is not only a form of robot, but also includes robot, language recognition, image recognition, natural language processing and expert system. Today, let's take a look at other products besides robots and artificial intelligence.
Google artificial intelligence project DeepMind
DeepMind, Google's research and development department in London, has developed artificial intelligence technology that can play video games independently. The computer system based on DeepMind technology can learn at an amazing speed, quickly master the gameplay and master the winning method of the game. Previously, the team called it a deep Q-network learning network. It only needs to observe the changes of the game screen and game score, and then it can analyze the game play and algorithm, and obtain "customs clearance skills" and high scores, which can reach the level of professional human players.
Google artificial intelligence project DeepMind
At present, under the same algorithm, network architecture and parameter settings, the system has been able to master 22 games (including the above-mentioned Space Invaders) after testing 49 games, reaching the level of expert games. This system further proves that artificial intelligence can master game skills through deep learning, gain the same control as human beings, and even surpass human beings in some aspects.
IBMWatson
Last year, IBM released WatsonAnalytics. WatsonAnalytics realizes the cognitive service based on natural language, which can provide real-time forecasting and visual analysis tools for business people. WatsonAnalytics will launch a free value-added application version based on cloud services at the end of this year, which can be used on computers and mobile devices.
IBMWatson
WatsonAnalytics can provide self-service analysis functions, including data access, data cleaning and data warehouse, which can help enterprise users to obtain and prepare data, analyze and visualize the results on this basis, and provide basis and convenience for users to take effective actions and further interaction.
Microsoft artificial intelligence torque Chinese version
In February this year, Microsoft released the Chinese version of Torque, which is an artificial intelligence product tailored for users of Android platform in China. It is gesture-driven and voice-interactive. As Microsoft's first artificial intelligence product on the Android platform, it is also Microsoft's first Chinese product for wearable devices. The goal of Torque is to transmit information in the most direct and timely way with the smallest interface. The birth of Torque liberated the hands of Android users. Users only need to gently shake their wrists, and then say "Happy Camp Host", "Where is the nearest KFC", "Call Zhang Yong" and other instructions, so they can easily obtain information and complete more tasks with minimal actions. This is exactly Microsoft's understanding of mobile productivity and efficiency in the era of mobile Internet.
Microsoft torque Chinese version
Microsoft said that Torque, like Xiao Bing, Xiao Na and other Microsoft artificial intelligence products, uses Bing big data platform as the underlying engine to process voice commands uploaded to the cloud by each user through mobile phones and mobile Internet. According to the preferences and habits of users in China, the artificial intelligence product team of Microsoft (Asia) Internet Engineering Institute has made special design and localized development in terms of functions.
Youtube subtitles
In 2009, Google used their speech recognition technology to provide a real-time "AutomaticCaptions" function on YouTube, which not only allows users to watch thousands of movies on the network without disturbing others and turning on the speakers.
Youtube subtitles
YouTube calls Google's automatic speech recognition technology (ASR) to add subtitles to YouTube videos. This technology comes from GoogleVoice. Of course, the generated subtitles can't be 100% accurate, but at least it can help you improve your listening and understand the video content, and Google will always improve its automatic speech recognition technology. This technology supports automatic subtitles in English, Japanese, Korean, Spanish, German, Italian, French, Portuguese, Russian and Dutch.
In addition to the automatic subtitle function, YouTube also adds subtitle time and automatic time function for friends who subtitle videos, making it easier for everyone to subtitle themselves. You just need to create a simple text file and write everything in the video, and then Google can use automatic speech recognition technology to match these words in the text with the words it recognizes, so that the accuracy is improved and it doesn't take too much time to subtitle sentence by sentence.
Artificial Intelligence bionic eye
The Royal Eye Hospital in Manchester, England, successfully performed the world's first artificial bionic robotic eye transplant to treat blindness caused by age-related macular degeneration (AMD). This artificial intelligence bionic eye device, named ArgusII, consists of two parts: the implanted part in vivo and the part that patients must wear in vitro. The implant device will be implanted into the retina of the patient, which contains an electrode array, a battery and a wireless antenna. External equipment includes a pair of glasses, a built-in front camera, a radio transmitter and a video processing unit.
Artificial Intelligence bionic eye
The camera will capture the picture directly opposite the implant and send the signal to the video processor for processing. The processed signal is sent back to the glasses, and the signal is received by the antenna of the implanted device through the glasses. Finally, the video is "output" to the electrode array, which plays the role of optic nerve simulation. The resolution of the electrode array reaches 60 pixels, which is enough for the implant device to track the trajectory of the object, clearly see the basic patterns and shapes, or slowly read larger characters. The pictures provided by ArgusII are black and white, but the development team of Argus is working hard to write electrode brain stimulation, hoping to make the brain receive color signals as soon as possible.
After the operation, the patient can recognize the vertical or horizontal lines and faces after recovery, and does not need a magnifying glass to read the newspaper. More interestingly, through this operation, patients can see the scene of their eyes even if they close their eyes, which makes people feel a little interesting.
In addition, the artificial intelligence eyeball developed by Second Vision, an American company that develops artificial intelligence eyeball, has also been approved for marketing, and this product can restore the vision of the completely blind.
News writing robot
News writing robot
The Associated Press used Wordsmith platform to write financial news automatically last summer. According to LouFerrara, the business news director of the Associated Press, after adopting the algorithm-based machine news writing, the amount of reports on the quarterly business situation of enterprises in the business news of the Associated Press will increase by more than 10 times, that is, from the original 300 to 4,400 articles per quarter, and at the same time, it will "liberate" journalists who used to report such news before, so that they can engage in more creative and challenging news. When the system is just online, it still needs manual review and adjustment of the platform. After three months, there is no need for human intervention
Bird face recognition technology developed by Cornell University
Cornell University and VIsipedia research project team jointly developed MerlinBirdPhotoID software, which can identify bird species in various pictures with the help of computer vision recognition technology and deep machine learning. For beginners and bird lovers, this is a very good software. Through deep machine learning, this program can provide recognition results in a few seconds, and the accuracy of the first three recognition results has reached more than 90%.
Cornell University develops bird facial recognition technology.
Users can frame the bird images that need to be identified by uploading bird images of unknown species, thus narrowing the identification range. The software can point out the known bird species from tens of thousands of pictures. At present, the database already contains more than 400 species of birds common in North America. With the number of users' use and deep machine learning, the accuracy will increase day by day. SergeBelongie, a professor at Cornell University, said: "Computers can process pictures more effectively than humans. They can classify, index and process a large number of graphic details to identify the results. "
Skype real-time translation tool
SkypeTranslator, Microsoft's real-time translation tool, combines speech recognition technology with what Microsoft calls "deep neural network and Microsoft's proven static machine translation technology". Can automatically translate voice calls and instant messages in different languages.
Currently, English, Spanish, Italian and Mandarin Chinese are supported. In addition, the translation of instant messaging messages has supported 50 languages, including French, Japanese, Arabic, Welsh and even Klingon.
High altitude translator
Because this translation tool integrates advanced technologies such as machine translation, speech recognition, machine learning and big data, it is widely optimistic. It is understood that the speech recognition technology for Skype Chinese real-time interpretation was jointly developed by Microsoft China and American researchers.