Shulou(Shulou.com)06/02 Report--

Author | James Murphy

Translator | Tiandao reward and diligence editor | Carol

Produced | AI Technology Camp (ID:rgznai100)

The subset of artificial intelligence (AI) has multiplied and accomplished a variety of tasks that only humans can accomplish. Techniques such as machine learning can perform administrative tasks, face recognition, chess, and even language translation. There is no doubt that the decade of artificial intelligence has brought many developments. In addition, deep learning learns from unstructured data to prepare analytical reports or perform unsupervised tasks. All these developments have laid the foundation for different companies to play a role and prove their value. As a result, many companies like DeepMind have been set up to continue to develop this field. What do you know about it? Let's take a look at everything about Google DeepMind today.

Google DeepMind's history DeepMind Technologies was founded in London in 2010, but four years later, Google acquired the company. Its ownership also changed in 2015 because it was later acquired by Alphabet,Inc, and it has been a subsidiary of the company ever since. DeepMind was originally founded by Demis Hassabis,Mustafa Suleyman and Shane Legg, both of whom are enthusiasts of artificial intelligence, and some see them as pioneers of deep learning. Since its inception, DeepMind Technologies has opened research centers in the United States, Canada and France. Since AlphaGo beat the world go champion Lee Sedol in 2016, it has been recognized by many people. The game was recorded, and after people saw it, they began to trust the company. In addition, they developed another program called AlphaZero, which can play chess and Japanese chess and achieved the best results. DeepMind has received considerable financial support with the participation of people such as Scott Scott Banister and Elon Musk. This is the extra capital they received from venture capital firms, Horizons Ventures (Horizons Ventures) and Founders Fund. The founders of DeepMind made reliable introductions to these entities, which is why they received the funding. If you also have an innovative and promising idea, it is wise to look up the best pitch board and create a highly professional and effective presentation. General learning algorithm DeepMind is very interesting in general learning algorithms. It can not only improve this field, but also help people better understand the human brain. The company has begun to achieve this goal by developing systems that can play a variety of games. One of the founders mentioned that they believe that artificial intelligence can reach the human level when a program can play a variety of different games. Their strategy is supported by scientific research that proves that games such as chess can improve strategic thinking. By learning how to play these complex games, machines will gain the ability to think and take strategic action. DeepMind's general learning algorithm allows machines to try to acquire humanoid intelligence and behavior through game-based learning. Although the company is very interested in machine learning to achieve human intelligence, it also has an objective view of the safety of using these technologies. To avoid machine disaster, DeepMind developed an open source test platform to determine whether the algorithm has a stop switch in the event of bad behavior. This open source test platform, called GridWorld, ensures that AI is safe and harmless to itself, developers, and others who have access to it.

DeepMind's deep reinforcement learning DeepMind promotes deep learning to a new level by implementing a completely different technology system. This system is called deep reinforcement learning, which is completely independent from the conventional artificial intelligence system. For example, IBM Watson or Deep Blue is developed for a specific purpose and is programmed to run at only the required capacity. DeepMind's deep reinforcement learning is not pre-programmed, but is learned through experience like human beings. In essence, it carries out deep learning based on convolution neural network and matches it with Q-learning. Their system was then tested on various video games without having to write instructions on how to play the game. Everything is done independently by the system. It learns how to play video games. After many attempts, it plays better than anyone else. The system has played all kinds of games and has mastered it better than those who are best at playing games. Deep reinforcement learning eliminates any human errors that may interfere with the efficiency of the game. It is used not only in games, but also in a variety of useful systems that have an impact on the health care industry.

WaveNet collaboration: speaking up for people with language disorders WaveNet collaboration is one of the most remarkable medical developments that DeepMind has contributed to. Millions of people suffer from language disorders and are unable to restore their original voices. Text-to-speech systems usually produce mechanical or unnatural sounds. DeepMind works with people with language disorders such as Google and Tim Tim Shaw, who has amyotrophic lateral sclerosis (ALS). The aim is to develop a system that sounds like a patient's natural voice, which at first glance seems impossible. Reproducing sound requires an individual to read hours of audio recording of a particular script. Unfortunately, people with language barriers may not have this luxury because they can't even easily form a sentence. DeepMind studied an algorithm that requires only a small number of recordings to reproduce sound. Six months later, the WaveNet collaboration has played a role in Tim's voice and presented it to him and his family. The results surprised them because it sounded like a voice before ALS began to affect Tim's voice capabilities. You can see people's reaction on YouTube because the whole process was filmed and uploaded.

Other contributions to Google DeepMind has been involved in a lot of development work, many of which are targeted at Google's artificial intelligence department. Personalized application advice is one of the most popular applications that most people use every day. DeepMind's artificial intelligence system collects data about your preferences and recommends apps similar to the ones you downloaded earlier. A more complex project they worked on was to create algorithms to cool the temperature of Google servers in the data center. DeepMind systems have improved the efficiency of these cooling systems, and Google has bigger plans for the company. Soon, users of Android Pie devices will have functions such as adaptive brightness and batteries. Machine learning will help these devices save energy by adjusting the brightness to adapt to the current lighting conditions. In addition, it will make the operating system easier to use, thereby improving the user experience. Because of the small scale of the project, creating these systems can be a bit complex. Such machine learning systems usually require more computing power to run successfully.

Key DeepMind has made great progress in the field of artificial intelligence and launched many useful innovative systems. Its contribution to Google's artificial intelligence department is valuable and has been used around the world. On the other hand, DeepMind also works with companies such as WaveNet to add value to people's lives. Because of the particularity of the artificial intelligence system they use, deep reinforcement learning makes them the company of choice for Google. What else do you expect DeepMind to do? Or after reading this article, what possibilities do you want to explore in artificial intelligence? You can discuss it with us in the comments section.

Original: https://hackernoon.com/all-we-need-to-know-about-googles-deepmind-0u6532r9

Https://mp.weixin.qq.com/s/EDNEVKC_HOp7wUEJeIcRJw

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.