artificial intelligence, transparency, the spread of viruses directly at the push of a button when and where it is required. it is a question of time before the population resists these interventions. there is no way out and no solution, only destruction
DeepMind (formerly Google DeepMind) is a company that specializes in programming artificial intelligence (AI).
DeepMind was founded in September 2010 and acquired by Google Inc. in 2014.
DeepMind Technologies was a British start-up founded in 2010 by Demis Hassabis, Shane Legg and Mustafa Suleyman. The first donors included the venture capital companies Horizons Ventures and Founders Fund and the business angel Scott Banister. In 2014 the company was awarded the „Company of the Year“ award by the Cambridge Computer Laboratory.
On January 26, 2014, the US company Google announced the takeover of DeepMind Technologies, thereby outperforming Facebook, which is also interested in DeepMind – it was the largest Californian takeover in Europe to date. The price remained secret, industry experts assumed a purchase price of about $ 500 million (365 million euros). Several billionaires had already invested in DeepMind, including Elon Musk, CEO of the private space company SpaceX and electric car manufacturer Tesla Motors, PayPal founder and original Facebook sponsor Peter Thiel, Skype co-founder Jaan Tallinn and Hong Kong magnate Li Ka-shing from Horizon Ventures , one of the most powerful men in Asia.
In 2014 DeepMind received the „Company of the Year“ award from the Cambridge Computer Laboratory.
After the takeover, the company was renamed Google DeepMind. Google also set up an Ethics Council to ensure that DeepMind’s technology is not misused. The structure of the Ethics Council remained unclear.
In 2015, the company published the results of a research project in which artificial intelligence was to learn old Atari games independently. According to Google DeepMind, the AI was able to both learn the rules of the game and develop successful tactics independently.
In December 2019, one of the founders, Suleyman, announced that he would leave DeepMind to work at Google.
The official corporate goal of Google DeepMind is to understand intelligence (“Solve Intelligence”). In contrast to other artificial intelligence such as Deep Blue from IBM, Google DeepMind has no predefined goal and is therefore more flexible in its use for various problems. Google DeepMind also differs in the basic structuring of artificial intelligence. Instead of relying solely on a neural network, the AI was expanded with a short-term memory in order to simulate the ability of an artificial memory. The developers of Google DeepMind therefore also call artificial intelligence a “neural Turing machine” and not a neural network.
DeepMind developed AlphaGo, a computer program that only plays the board game Go. In October 2015 it defeated the multiple European champion Fan Hui. It is the first program to beat a professional go player under tournament conditions. Between March 9 and 15, 2016, AlphaGo competed against South Korean professional Lee Sedol, 9th Dan. The program won 4-1 after five rounds.
In 2017, an improved version called AlphaGo Zero was released, which beat AlphaGo 100 to 0, with AlphaGo Zero’s strategies created autonomously. The learning phase lasted only three days, whereby AlphaGo took months in comparison.
In January 2019, AlphaStar was launched, an AI program that plays the real-time strategy game StarCraft II. Like AlphaGo, this is an artificial neural network that first imitated human players and was then trained with reinforcement learning. AlphaStar won every game in two sets of five games against professional players Dario „TLO“ Wünsch and Grzegorz „MaNa“ Komincz. The number of actions per minute has been limited to what is common for humans. In contrast to human players, AlphaStar always has a complete overview of the visible parts of the map, but still focuses on individual areas. MaNa won a game broadcast live during the performance. AlphaStar’s particular strengths were micromanagement and multitasking. In the last game, she was knocked out by unusual human player actions.
WaveNet is a neural network that was originally developed from around 2014 for the purpose of improved, more natural-sounding text-to-speech synthesis (TTS), that is, for human-sounding reading systems by which the program was supposed to analyze natural human voices, and so on to be able to produce natural-sounding speech when reading texts. Google has been offering the Cloud Text-to-Speech reading program based on WaveNet since March 2018 as part of the Google Assistant.
In the meantime, DeepMind is also actively researching to use WaveNet to reproduce existing individual human voices as precisely as possible, not only for reading out texts, in order to provide it with functions similar to those of Adobe Voco, which was launched by Adobe Inc. in 2016. A June 2018 research report entitled Disentangled Sequential Autoencoder found that WaveNet had successfully used to replace the voice in an existing sound recording with any other real voice, that speaks the same text („content swapping“). To enable this ability, about 50 hours of recordings of the source and target voice are required, with which WaveNet first has to learn the basic properties of the two voices (whereby it independently has its own model of the voice to be learned, abstracted from the voice recordings before it can do the voice conversion on any new sound recording of acceptable quality. The authors of the research report also emphasize that WaveNet is able to differentiate between static and dynamic properties (to: „unlink“); H. the program automatically separates the properties to be retained during voice conversion (text content, modulation, speed, mood, etc.) and the basic properties of the source and target voice to be converted.
In a follow-up report entitled Unsupervised speech representation learning using WaveNet autoencoders („Independent language model learning with WaveNet“) from January 2019, it was said that DeepMind had further improved the distinction between static and dynamic voice characteristics of WaveNet. In the further follow-up report, Sample Efficient Adaptive Text-to-Speech („Sampling Efficiency with Adaptive Text-to-Speech Synthesis“) from September 2018 (last revision January 2019) DeepMind reports the minimum required voice recordings for sampling one I have now reduced my voice to a few minutes.
In the Towards End-to-End Speech Recognition with Recurrent Neural Networks report, published in 2014, DeepMind referred to successful attempts to use WaveNet to automatically spell existing voice recordings.
The research report Large-Scale Visual Speech Recognition („General-purpose optical speech recognition“) from July 2018 (last revision from October 2018) deals with successful attempts to use WaveNet for lip reading in human lip movements even in completely silent video recordings Program already clearly surpass professional human lip readers when guessing the actual sounds. This feature seems to be developed as an independent WaveNet plugin under the title LipNet.
In February 2016, the company announced that it was working with the National Health Service in the UK to develop an iPhone app called Streams, but it was too early to say where AI would be used could. This is to help monitor patients suffering from „kidney damage“. In April 2016, the specialist magazine „New Scientist“ published details of the agreement, according to which Google will have access to the data of 1.6 million NHS policyholders who visit the Royal Free NHS ‚three hospitals – Barnet, Chase Farm and the Royal Free – each year Trust are treated. The far-reaching agreement allows data to be exchanged far beyond what was announced. It contains data from HIV-positive patients, drug addicts or women who have had abortions. There was no opt-out option (possibility of refusal) for patients. In addition to the pathological and radiological results, the data from intensive care medicine and the emergency departments are transmitted, as well as the full daily activities of the clinics, the condition and accommodation of the patients and hospital visits (who and when) are transmitted. In addition, access to data on the central records of all NHS hospital treatments in the UK means that it can access historical data from the past five years, in addition to the new data.
Critics fear privacy and data protection. However, Google staff should not be able to identify specific patients and the data should not be linked to Google accounts or products, said Mustafa Suleyman, head of „Applied AI“ at DeepMind.
DeepMind Ethics and Society
In October 2017, DeepMind announced that it would establish the DeepMind Ethics & Society research department, which will deal with the following topics: privacy, transparency and justice, and the economic consequences thereof.