Changes

Jump to: navigation, search

Timeline of DeepMind

8,669 bytes added, 12:01, 5 July 2019
no edit summary
| 2010–2014 || DeepMind initiates as a British AI startup. Before being acquired by {{w|Google}}, it remains relatively unknown.<ref>{{cite web |title=How Google's Amazing AI Start-Up 'DeepMind' Is Making Our World A Smarter Place |url=https://www.forbes.com/sites/bernardmarr/2017/02/02/how-googles-amazing-ai-start-up-deepmind-is-making-our-world-a-smarter-place/#43263b6edfff |website=forbes.com |accessdate=1 June 2019}}</ref>
|-
| Since 2014 || Google's DeepMind era. The acquired company starts being known worldwide. Since being acquired by Google, DeepMind's AI would be used to beat humans at board games and create free apps with the British {{w|National Health Service}}. Neither application would make profit for Google so far.<ref name="Google's $500+ million purchase of DeepMind just got very interesting"/>
|-
| Since 2016 || DeepMind becomes renowned after its {{W|AlphaGo}} program beats a human professional Go player for the first time and again when AlphaGo beats {{W|Lee Sedol}}, the world champion, in a five-game match.<ref>{{cite web |title=DeepMind |url=https://paganresearch.io/details/deepmind |website=paganresearch.io |accessdate=5 June 2019}}</ref>
| 2009 || || Prelude || {{w|Demis Hassabis}} attains a doctorate in {{w|Cognitive Neuroscience}} from {{w|University College London}}.<ref name="Google buys UK artificial intelligence startup Deepmind for £400m"/>
|-
| 2010 || September || Creation || {{w|DeepMind Technologies}} is co-founded by {{w|Demis Hassabis}} alongside {{w|Shane Legg}}, a machine learning researcher from {{w|New Zealand}}, and childhood friend {{w|Mustafa Suleyman}}.<ref name="Google DeepMind: the story behind the world's leading AI startup"/><ref>{{cite web |title=DeepMind’s elusive third cofounder is the man making sure that machines stay on our side |url=https://www.businessinsider.com/shane-legg-google-deepmind-third-cofounder-artificial-intelligence-2017-1 |website=businessinsider.com |accessdate=5 June 2019}}</ref><ref>{{cite web |title=Neuroscience, intuition and superhumans - how DeepMind co-founder and UCL alumnus Demis Hassabis is leading the Artificial Intelligence revolution |url=https://www.ucl.ac.uk/news/2016/nov/neuroscience-intuition-and-superhumans-how-deepmind-co-founder-and-ucl-alumnus-demis |website=ucl.ac.uk |accessdate=5 June 2019}}</ref>
|-
| 2011 || February 1 || Funding || DeepMind raises an undisclosed amount from {{w|Founders Fund}} and {{w|Horizons Ventures}}.<ref name="DeepMind Recent News & Activity">{{cite web |title=DeepMind > Recent News & Activity |url=https://www.crunchbase.com/organization/deepmind/timeline/timeline#section-recent-news-activity |website=crunchbase.com |accessdate=29 May 2019}}</ref>
|-
| 2012 || December || Acquisition || {{w|Facebook}} reportedly fails to acquire DeepMind, according to {{w|Forbes}}.<ref name="Google buys UK artificial intelligence startup Deepmind for £400m">{{cite web |title=Google buys UK artificial intelligence startup Deepmind for £400m |url=https://www.theguardian.com/technology/2014/jan/27/google-acquires-uk-artificial-intelligence-startup-deepmind |website=theguardian.com |accessdate=29 May 2019}}</ref> Discussions between DeepMind talks with and Facebook about the acquisition happen around the this time.<ref name="Google DeepMind"/>
|-
| 2013 || December || {{w|AI}} development || DeepMind reveals having developed an AI algorithm able to learn how to play iconic early video games like [[w:Breakout (video game)|Breakout]] and {{w|Pong}} simply by watching them being played on a vintage 1977 Atari 2600 games console. The algorithm deduces the rules and rewards from the way the pixels are batted about the screen, then it is able to beat human opponents at playing the games.<ref name="Google buys AI firm DeepMind to boost image search">{{cite web |title=Google buys AI firm DeepMind to boost image search |url=https://www.newscientist.com/article/dn24946-google-buys-ai-firm-deepmind-to-boost-image-search/ |website=newscientist.com |accessdate=28 May 2019}}</ref>
|-
| 2013 || December || Team || DeepMind reportedly has around 75 employees.<ref name="Google DeepMind">{{cite web |title=Google To Acquire Artificial Intelligence Company DeepMind |url=https://www.forbes.com/sites/amitchowdhry/2014/01/27/google-to-acquire-artificial-intelligence-company-deepmind/#2d9358fa15b5 |website=forbes.com |accessdate=29 May 2019}}</ref>
|-
| 2013 || || Team || [[w:David Silver (programmer)|David Silver]] joins DeepMind as a full-time research scientist.<ref>{{cite web |title=Computer Science News |url=http://www.cs.ucl.ac.uk/news/article/david_silver_of_deepmind_delivers_inaugural_lecture_at_ucl/ |website=cs.ucl.ac.uk |accessdate=5 July 2019}}</ref>
|-
| 2014 || January 27 || Acquisition || {{w|Google}} engages in its largest European acquisition to date with a deal to buy DeepMind for £400 million (US$650 million). From then on, DeepMind starts becoming known worldwide.<ref name="Google's $500+ million purchase of DeepMind just got very interesting">{{cite web |last1=Shead |first1=Sam |title=Google's $500+ million purchase of DeepMind just got very interesting |url=https://www.businessinsider.com/googles-400-million-acquisition-of-deepmind-is-looking-good-2016-7 |website=businessinsider.com |accessdate=5 June 2019}}</ref><ref>{{cite web |title=Google buys UK artificial intelligence startup Deepmind for £400m |url=https://www.theguardian.com/technology/2014/jan/27/google-acquires-uk-artificial-intelligence-startup-deepmind |website=theguardian.com |accessdate=5 June 2019}}</ref><ref name="Google DeepMind: the story behind the world's leading AI startup">{{cite web |title=Google DeepMind: the story behind the world's leading AI startup |url=https://www.techworld.com/startups/google-deepmind-what-is-it-how-it-works-should-you-be-scared-3615354/ |website=techworld.com |accessdate=28 May 2019}}</ref><ref name="Google buys AI firm DeepMind to boost image search"/><ref>{{cite web |title=Google acquires artificial intelligence company DeepMind |url=https://www.computerworld.com/article/2486983/google-acquires-artificial-intelligence-company-deepmind.html |website=computerworld.com |accessdate=29 May 2019}}</ref>
|-
| 2014 || October 29 || {{w|AI}} development || DeepMind unveils a neural network that can access an external memory like a conventional {{w|Turing machine}}. The project mimics properties of the human brain's short-term working memory. The result is a computer able to mimic some of the brain’s memory skills and even program like a human.<ref>{{cite web |title=Google's Secretive DeepMind Startup Unveils a "Neural Turing Machine" |url=https://www.technologyreview.com/s/532156/googles-secretive-deepmind-startup-unveils-a-neural-turing-machine/ |website=technologyreview.com |accessdate=29 May 2019}}</ref><ref>{{cite web |last1=Hornyak |first1=Tim |title=Google's DeepMind AI project apes human memory and programming skills |url=https://www.pcworld.com/article/2841232/google-ai-project-apes-memory-programs-sort-of-like-a-human.html |website=pcworld.com |accessdate=29 May 2019}}</ref><ref>{{cite web |title=Computer with human-like learning will program itself |url=https://www.newscientist.com/article/mg22429932-200-computer-with-human-like-learning-will-program-itself/ |website=newscientist.com |accessdate=29 May 2019}}</ref><ref>{{cite web |title=Google reveals it is developing a computer so smart it can program ITSELF |url=https://www.dailymail.co.uk/sciencetech/article-2813303/Google-reveals-developing-computer-smart-program-ITSELF.html |website=dailymail.co.uk |accessdate=29 May 2019}}</ref>
|-
| 2014 || November || Notable comment || {{w|Elon Musk}} submits a comment to <code>edge.com</code> about the threat of AI:
 
{{Quote|The pace of progress in artificial intelligence (I'm not referring to narrow AI) is incredibly fast. Unless you have direct exposure to groups like Deepmind, you have no idea how fast-it is growing at a pace close to exponential. The risk of something seriously dangerous happening is in the five year timeframe. 10 years at most. This is not a case of crying wolf about something I don't understand.
 
I am not alone in thinking we should be worried. The leading AI companies have taken great steps to ensure safety. The recognize the danger, but believe that they can shape and control the digital superintelligences and prevent bad ones from escaping into the Internet. That remains to be seen...}}<ref>{{cite web |title=28 Best Quotes About Artificial Intelligence |url=https://www.bernardmarr.com/default.asp?contentID=1158 |website=bernardmarr.com |accessdate=29 June 2019}}</ref>
|-
| 2014 || || Recognition || DeepMind receives the "Company of the Year" award from {{w|Cambridge Computer Laboratory}}.<ref>{{cite web|title=Hall of Fame Awards: To celebrate the success of companies founded by Computer Laboratory graduates.|url=https://www.cl.cam.ac.uk/ring/awards.html|publisher=University of Cambridge|accessdate=28 May 2019}}</ref>
|-
| 2015 || June || Notable quote comment || {{w|Mustafa Suleyman}} describes DeepMind's work during a {{w|machine learning}} conference in {{w|London}}: {{Quote|"Our deep learning tool has now been deployed in many environments, particularly across Google in many of our production systems."}}<ref name="Google DeepMind: the story behind the world's leading AI startup"/>|-| 2015 || June || Team || [[w:Alex Graves (computer scientist)|Alex Graves]] joins DeepMind as a research scientist.<ref name="The 21 smartest AI scientists working at Google DeepMind">{{cite web |title=The 21 smartest AI scientists working at Google DeepMind |url=https://www.businessinsider.com/the-smartest-ai-scientists-working-at-google-deepmind-2015-11 |website=businessinsider.com |accessdate=5 July 2019}}</ref>
|-
| 2015 || September || Partnership || DeepMind partners with the Royal Free NHS Trust to develop a patient safety app called Streams, application aimed at reviewing test results for signs of sickness and sending staff instant alerts if an urgent assessment is required. The app would also help clinicians to quickly check for other serious conditions such as acute kidney injury and display results of blood tests, scans, and x-rays at the touch of a button.<ref name="Google DeepMind: the story behind the world's leading AI startup"/><ref>{{Cite news|url=https://techcrunch.com/2017/08/31/documents-detail-deepminds-plan-to-apply-ai-to-nhs-data-in-2015/|title=Documents detail DeepMind’s plan to apply AI to NHS data in 2015|last=Lomas|first=Natasha|work=TechCrunch|access-date=28 May 2019|language=en}}</ref>
|-
| 2015 || October || Achievement || DeepMind's {{w|AlphaGo}} beats the European [[w:Go (game)|Go]] champion {{w|Fan Hui}}, a [[w:Go ranks and ratings|2 dan]] (out of 9 dan possible) professional, five to zero.<ref name="bbcgo">{{cite news|url=https://www.bbc.com/news/technology-35420579|title=Google achieves AI 'breakthrough' by beating Go champion|date=27 January 2016|author=|newspaper=BBC News}}</ref> This is the first time an {{w|artificial intelligence}} defeats a professional Go player.<ref name="lemondego">{{cite news|url=http://www.lemonde.fr/pixels/article/2016/01/27/premiere-defaite-d-un-professionnel-du-go-contre-une-intelligence-artificielle_4854886_4408996.html|title=Première défaite d'un professionnel du go contre une intelligence artificielle|language=French|date=27 January 2016|work=Le Monde}}</ref>
|-
| 2015 || February 25 || {{w|AI}} development || DeepMind builds develops an {{w|artificial intelligence agent that can learn }} capable of learning how to successfully play 49 classic {{w|Atari games}} by itself, with minimal input. The researchers claim software that learns to play video games could graduate to the real world before long.<ref>{{cite web |last1=Twilley |first1=Nicola |title=Artificial Intelligence Goes to the Arcade |url=https://www.newyorker.com/tech/annals-of-technology/deepmind-artificial-intelligence-video-games |website=newyorker.com |accessdate=29 May 2019}}</ref><ref>{{cite web |last1=Clark |first1=Liat |title=DeepMind's AI is an Atari gaming pro now |url=https://www.wired.co.uk/article/google-deepmind-atari |website=wired.co.uk |accessdate=29 May 2019}}</ref><ref>{{cite web |title=Google's AI Masters Space Invaders (But It Still Stinks at Pac-Man) |url=https://www.technologyreview.com/s/535446/googles-ai-masters-space-invaders-but-it-still-stinks-at-pac-man/ |website=technologyreview.com |accessdate=29 May 2019}}</ref><ref>{{cite web |last1=Devlin |first1=Hannah |title=Google develops computer program capable of learning tasks independently |url=https://www.theguardian.com/technology/2015/feb/25/google-develops-computer-program-capable-of-learning-tasks-independently |website=theguardian.com |accessdate=29 May 2019}}</ref>
|-
| 2016 || February || Partnership || DeepMind announces that it is teaming with the {{w|National Health Service}} to build an app called Streams to help hospital staff monitor patients with kidney disease.<ref name="newscientist.com">{{cite web |title=Revealed: Google AI has access to huge haul of NHS patient data |url=https://www.newscientist.com/article/2086454-revealed-google-ai-has-access-to-huge-haul-of-nhs-patient-data/ |website=newscientist.com |accessdate=28 May 2019}}</ref>
| 2016 || April || Controversy || ''{{w|New Scientist}}'' obtains a copy of a data-sharing agreement between DeepMind and the {{w|Royal Free London NHS Foundation Trust}}. The latter operates three London hospitals where an estimated 1.6 million patients are treated annually. The agreement shows DeepMind Health had access to admissions, discharge and transfer data, accident and emergency, pathology and radiology, and critical care at these hospitals, including personal details such as whether patients have been diagnosed with [[w:HIV/AIDS|HIV]], suffered from [[w:major depressive disorder|depression]] or have ever undergone an {{w|abortion}} in order to conduct research to seek better outcomes in various health conditions.<ref>{{cite news |url=https://www.newscientist.com/article/2086454-revealed-google-ai-has-access-to-huge-haul-of-nhs-patient-data |title=Revealed: Google AI has access to huge haul of NHS patient data |first=Hal |last=Hodson |work={{w|New Scientist}} |date=29 April 2016 |accessdate= 28 May 2019}}</ref><ref>{{cite news |url=https://www.newscientist.com/article/mg23030722-900-big-data-if-theres-nothing-to-hide-why-be-secretive/ |title=Leader: If Google has nothing to hide about NHS data, why so secretive? |work={{w|New Scientist}} |date=4 May 2016 |accessdate= 28 May 2019}}</ref>
|-
| 2016 || June 3 || {{w|AI}} development || DeepMind develops a ‘big red button’ to stop AIs from causing harm, using a framework in the form of "safely interruptible" artificial intelligence. The system It guarantees that a machine will not learn to resist attempts by humans to intervene in its learning processes.<ref>{{cite web |last1=Byrne |first1=Michael |title=Google DeepMind Researchers Develop AI Kill Switch |url=https://www.vice.com/en_us/article/bmv7x5/google-researchers-have-come-up-with-an-ai-kill-switch |website=vice.com |accessdate=30 May 2019}}</ref><ref>{{cite web |last1=Li |first1=Abner |title=Google DeepMind has developed a ‘big red button’ to stop AIs from causing harm |url=https://9to5google.com/2016/06/03/google-deepmind-big-red-button-ai/ |website=9to5google.com |accessdate=30 May 2019}}</ref><ref>{{cite web |last1=Shead |first1=Sam |title=Google has developed a 'big red button' that can be used to interrupt artificial intelligence and stop it from causing harm |url=https://www.businessinsider.com/google-deepmind-develops-a-big-red-button-to-stop-dangerous-ais-causing-harm-2016-6 |website=businessinsider.com |accessdate=30 May 2019}}</ref>. The system is described in a paper by Laurent Orseau from DeepMind and Stuart Armstrong from the {{w|Future of Humanity Institute}}.<ref>{{cite web |title=Safely Interruptible Agents |url=https://intelligence.org/files/Interruptibility.pdf |website=intelligence.org |accessdate=11 June 2019}}</ref>
|-
| 2016 || July || {{w|AI}} development || DeepMind announces the ability to cut Google's data centers' energy consumption by 15%, using a {{w|machine learning}} algorithm.<ref>{{cite web |last1=Burgess |first1=Matt |title=Google's DeepMind trains AI to cut its energy bills by 40% |url=https://www.wired.co.uk/article/google-deepmind-data-centres-efficiency |website=wired.co.uk |accessdate=29 May 2019}}</ref><ref>{{cite web |last1=Wakefield |first1=Jane |title=Google uses AI to save on electricity from data centres |url=https://www.bbc.com/news/technology-36845978 |website=bbc.com |accessdate=29 May 2019}}</ref>
|-
| 2016 || December 5 || Userbase || DeepMind announces open-sourcing DeepMind Lab, its 3D game-like platform for agent-based AI research, so that others can try and make advances in the field of AI. The DeepMind Lab project was used to create enviroments capable of testing AI systems’ ability to achieve goals in a wide range of environments. Tasks such as navigation in mazes, collecting fruit, traversing dangerous passages, laser tag and interaction with bots have been developed to refine the programs. The development of mazes and challenges were designed using video game {{w|Quake III Arena}}’s 17-year-old software, to teach its artificial intelligence programs how to operate in 3D spaces.<ref>{{cite web |last1=McCarthy |first1=John |title=Google DeepMind releases source code to the Quake III levels its using to train AIs |url=https://www.thedrum.com/news/2016/12/05/google-deepmind-releases-source-code-the-quake-iii-levels-its-using-train-ais |website=thedrum.com |accessdate=30 May 2019}}</ref><ref>{{cite web |last1=Condon |first1=Stephanie |title=OpenAI, DeepMind open source AI training platforms |url=https://www.zdnet.com/article/openai-deepmind-open-source-ai-training-platforms/ |website=zdnet.com |accessdate=30 May 2019}}</ref><ref>{{cite web |last1=Shead |first1=Sam |title=DeepMind is opening up its 'flagship' platform to AI researchers outside the company |url=https://www.businessinsider.com/deepmind-opens-up-lab-to-ai-researchers-2016-12 |website=businessinsider.com |accessdate=30 May 2019}}</ref><ref>{{cite web |last1=Kahn |first1=Jeremy |title=Google DeepMind Makes AI Training Platform Publicly Available |url=https://www.bloomberg.com/news/articles/2016-12-05/google-deepmind-makes-ai-training-platform-publicly-available |website=bloomberg.com |accessdate=30 May 2019}}</ref>
|-
| 2016 || December || International expansion || DeepMind Applied is anounced as a sub-team to be forming in {{w|Mountain View}}, {{w|California}}. This team is expected to be more closely involved with {{w|Google}}, working with the various product teams to help implement AI-based solutions.<ref>{{cite web |title=New DeepMind Applied Team Being Formed In Mountain View |url=https://www.androidheadlines.com/2016/12/new-deepmind-applied-team-being-formed-in-mountain-view.html |website=androidheadlines.com |accessdate=5 July 2019}}</ref>
|-
| 2016 || || Financial || DeepMind records £40.3 million (US$ 52 million) in revenue in the year.<ref name="DeepMind Losses Grew To $368 Million In 2017"/>
|-
| 2017 || January || Collaboration || DeepMind's experts pledge to pass on their knowledge to students enrolled on machine learning master's programs at {{w|University College London}}.<ref>{{cite web |last1=Shead |first1=Sam |title=DeepMind's AI experts have pledged to pass on their knowledge to students at UCL |url=https://www.businessinsider.com/deepmind-ai-experts-are-going-to-teach-at-ucl-2017-1 |website=businessinsider.com |accessdate=1 June 2019}}</ref>
|-
| 2017 || May || Controversy || ''Sky News'' publishes a leaked letter from the National Data Guardian, [[wikipedia:Fiona Caldicott|Dame Fiona Caldicott]], revealing that in her "considered opinion" the data-sharing agreement between DeepMind and the Royal Free took place on an "inappropriate legal basis".<ref>{{cite news |url=http://news.sky.com/story/google-received-16-million-nhs-patients-data-on-an-inappropriate-legal-basis-10879142/ |title=Google received 1.6 million NHS patients' data on an 'inappropriate legal basis' |first=Alexander J |last=Martin |work={{w|Sky News}} |date=15 May 2017 |accessdate=28 May 2019}}</ref>
|-
| 2017 || June || Partnership || DeepMind’s safety team partners with {{w|OpenAI}} in the development of an algorithm which can infer what humans want by being told which of two proposed behaviors is better. The learning algorithm uses small amounts of human feedback to solve modern {{w|reinforcement learning}} environments.<ref>{{cite web |title=Learning from Human Preferences |url=https://openai.com/blog/deep-reinforcement-learning-from-human-preferences/ |website=openai.com |accessdate=29 June 2019}}</ref>
|-
| 2017 || July || International expansion || DeepMind announces its first international research lab in {{w|Edmonton}}, Canada.<ref>{{cite web |last1=Vincent |first1=James |title=Google’s AI powerhouse DeepMind is opening its first international lab in Canada |url=https://www.theverge.com/2017/7/5/15922574/deepmind-alberta-canda-research-lab |website=theverge.com |accessdate=31 May 2019}}</ref><ref>{{cite web |last1=Fingas |first1=Jon |title=Google's next DeepMind AI research lab opens in Canada |url=https://www.engadget.com/2017/07/05/google-opens-deepmind-lab-in-canada/ |website=engadget.com |accessdate=31 May 2019}}</ref><ref>{{cite web |last1=Kahn |first1=Jeremy |title=DeepMind Goes to Alberta For First International Lab |url=https://www.bloomberg.com/news/articles/2017-07-05/deepmind-goes-to-alberta-for-first-international-lab |website=bloomberg.com |accessdate=31 May 2019}}</ref><ref>{{cite web |last1=Semeniuk |first1=Ivan |title=AI company that conquered Go game opens office in Edmonton |url=https://www.theglobeandmail.com/news/national/ai-company-that-conquered-go-game-opens-office-in-edmonton/article35551899/ |website=theglobeandmail.com |accessdate=31 May 2019}}</ref>
| 2017 || October 4 || Team || DeepMind launches DeepMind Ethics & Society (DMES), a new research group recruiting advisers from academia and charity sector with the purpose to ‘help technologists put ethics into practice’ and help coping with artificial intelligence to consider the “real-world impacts” of replicating human intelligence. The group consists of six independent research fellows, eight full-time researchers, and nine partnerships with other research institutions. It would explore topics such as algorithmic bias, accountability, and autonomous killing machines.<ref>{{cite web |last1=Gershgorn |first1=Dave |title=DeepMind now has two ethics groups, but one of them is still secret |url=https://qz.com/1094283/alphabets-goog-deepmind-now-has-two-ethics-groups-but-one-of-them-is-still-secret/ |website=qz.com |accessdate=31 May 2019}}</ref><ref>{{cite web |last1=Hern |first1=Alex |title=DeepMind announces ethics group to focus on problems of AI |url=https://www.theguardian.com/technology/2017/oct/04/google-deepmind-ai-artificial-intelligence-ethics-group-problems |website=theguardian.com |accessdate=31 May 2019}}</ref><ref>{{cite web |last1=Sawers |first1=Paul |title=Alphabet’s DeepMind sets up ‘ethics and society’ unit to research real-world impact of AI |url=https://venturebeat.com/2017/10/04/alphabets-deepmind-sets-up-ethics-and-society-unit-to-research-real-world-impact-of-ai/ |website=venturebeat.com |accessdate=31 May 2019}}</ref>
|-
| 2017 || October 18 || {{w|AI}} development || DeepMind announces AlphaGo Zero, a software capable of mastering the Chinese game of {{w|Go}} without help from human players.The new version is an improvement on the original AlphaGo, which had to be trained over time using large quantities of human knowledge and supervision.<ref>{{cite web |title=AI versus AI: Self-Taught AlphaGo Zero Vanquishes Its Predecessor |url=https://www.scientificamerican.com/article/ai-versus-ai-self-taught-alphago-zero-vanquishes-its-predecessor/?redirect=1 |website=scientificamerican.com |accessdate=11 June 2019}}</ref><ref name="Google DeepMind: AI becomes more alien">{{cite web |last1=Cellan-Jones |first1=Rory |title=Google DeepMind: AI becomes more alien |url=https://www.bbc.com/news/technology-41668701 |website=bbc.com |accessdate=30 May 2019}}</ref><ref>{{cite web |last1=Burgess |first1=Matt |title=DeepMind's latest AI breakthrough is its most significant yet |url=https://www.wired.co.uk/article/deepmind-alphago-zero-nature-reinforcement-learning |website=wired.co.uk |accessdate=30 May 2019}}</ref><ref>{{cite web |last1=Cuthbertson |first1=Anthony |title=Deepmind AlphaGo: AI Teaches Itself 'Thousands of Years of Human Knowledge' Without Help |url=https://www.newsweek.com/deepmind-alphago-ai-teaches-human-help-687620 |website=newsweek.com |accessdate=30 May 2019}}</ref><ref>{{cite web |last1=Shead |first1=Sam |title=DeepMind's human-bashing AlphaGo AI is now even stronger |url=https://www.businessinsider.com/deepminds-alphago-ai-gets-alphago-zero-upgrade-2017-10 |website=businessinsider.com |accessdate=30 May 2019}}</ref>
|-
| 2017 || October 20 || Recognition || The DeepMind team behind AlphaGo is awarded the inaugural {{w|Marvin Minsky Medal}} by the {{w|International Joint Conference On Artificial Intelligence}} (IJCAI) in {{w|Stockholm}}, for outstanding achievements in the field of AI.<ref>{{cite web |last1=Gorey |first1=Colm |title=DeepMind team behind AlphaGo wins inaugural ‘Nobel Prize for AI’ |url=https://www.siliconrepublic.com/machines/deepmind-alphago-marvin-minsky-award |website=siliconrepublic.com |accessdate=1 June 2019}}</ref><ref>{{cite web |title=DeepMind team behind AlphaGo wins inaugural ‘Nobel Prize for AI’ |url=http://www.lionra.ie/feed-items/deepmind-team-behind-alphago-wins-inaugural-nobel-prize-for-ai/ |website=lionra.ie |accessdate=1 June 2019}}</ref><ref>{{cite web |title=IJCAI 2018 Kicks Off; DeepMind AlphaGo Wins Marvin Minsky Medal |url=https://medium.com/syncedreview/ijcai-2018-kicks-off-deepmind-alphago-wins-marvin-minsky-medal-56ff073f2c38 |website=medium.com |accessdate=1 June 2019}}</ref><ref>{{cite web |title=DeepMind wins the Minsky medal for AlphaGo |url=https://www.celi.it/en/blog/2017/10/deepmind-alphago-team-receive-inaugural-ijcai-marvin-minsky-medal/ |website=celi.it |accessdate=1 June 2019}}</ref>
|-
| 2017 || October || International expansion || DeepMind opens new research office in {{w|Montreal}}, led by {{w|McGill University}} professor {{w|Doina Precup}}.<ref>{{cite web |title=DeepMind opens new AI research office in Montreal |url=https://www.linkedin.com/pulse/deepmind-opens-new-research-office-montreal-andres-restrepo/ |website=linkedin.com |accessdate=5 July 2019}}</ref>
|-
| 2017 || November || Partnership || DeepMind announces a research partnership with the {{w|Cancer Research UK}} Center at {{w|Imperial College London}} with the goal of improving {{w|breast cancer}} detection by applying {{w|machine learning}} to {{w|mammography}}.<ref>{{cite web|url=https://siliconangle.com/blog/2017/11/24/google-deepmind-announces-new-research-partnership-fight-breast-cancer-ai/|title=Google DeepMind announces new research partnership to fight breast cancer with AI|date=24 November 2017|website=Silicon Angle}}</ref>
|-
| 2017 || December 5 || {{w|AI}} development || DeepMind team introduces AlphaZero, a program using generalized AlphaGo Zero's approach, which achieved within 24 hours a superhuman level of play in {{w|chess}}, {{w|shogi}}, and [[w:Go (game)|Go]], defeating world-champion programs, [[w:Stockfish (chess)|Stockfish]], [[w:Elmo (shogi engine)|Elmo]], and 3-day version of AlphaGo Zero in each case.<ref>{{Cite web|first1=David|last1= Silver|first2=Thomas|last2= Hubert|first3= Julian|last3=Schrittwieser|first4= Ioannis|last4=Antonoglou |first5= Matthew|last5= Lai|first6= Arthur|last6= Guez|first7= Marc|last7= Lanctot|first8= Laurent|last8= Sifre|first9= Dharshan|last9= Kumaran|first10= Thore|last10= Graepel|first11= Timothy|last11= Lillicrap|first12= Karen|last12= Simonyan|first13=Demis |last13=Hassabis|title=Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm|class=cs.AI|date=5 December 2017|url=https://arxiv.org/abs/1712.01815}}</ref>
|-
| 2017 || || Financial || DeepMind records £54.4 million (US$ 71 million) in revenue in 2017, up 35% from £40.3 million (US$ 52 million) in 2016.<ref name="DeepMind Losses Grew To $368 Million In 2017">{{cite web |title=DeepMind Losses Grew To $368 Million In 2017 |url=https://www.forbes.com/sites/samshead/2018/10/05/deepmind-losses-grew-to-302-million-in-2017/#7719dff2490e |website=forbes.com |accessdate=29 June 2019}}</ref>
|-
| 2018 || February || Partnership || DeepMind announces that it is teaming with the [[w:United States Department of Veterans Affairs|U.S. Department of Veterans Affairs]] in an attempt to use machine learning to predict the onset of acute kidney injury in patients, and also more broadly the general deterioration of patients during a hospital stay so that doctors and nurses can more quickly treat patients in need.<ref>{{cite web|url=https://venturebeat.com/2018/02/22/googles-deepmind-wants-ai-to-spot-kidney-injuries/|title=Google's DeepMind wants AI to spot kidney injuries|date=22 February 2018|website=Venture Beat}}</ref>
|-
| 2018 || February || {{w|AI}} development || DeepMind develops an artificial intelligence able to pass capable of an important psychological assessment ability that most children only develop the skills to pass at around age 4, which is to infer what someone else is thinking. This new technology is thought to have useful application in the future, from warfare to elderly care.<ref>{{cite web |last1=Revell |first1=Timothy |title=DeepMind AI is learning to understand the ‘thoughts’ of others |url=https://www.newscientist.com/article/mg23731673-400-deepmind-ai-is-learning-to-understand-the-thoughts-of-others/ |website=newscientist.com |accessdate=3 June 2019}}</ref><ref>{{cite web |title=Will A.I. Ever Be Smarter Than a Four-Year-Old? |url=https://www.smithsonianmag.com/innovation/will-ai-ever-be-smarter-than-four-year-old-180971259/ |website=smithsonianmag.com |accessdate=3 June 2019}}</ref><ref>{{cite web |title=‘Deep Mind’ AI As Advanced as a 4yr old! |url=https://www.trebuchet-magazine.com/deep-mind-ai-advanced-4yr-old/ |website=trebuchet-magazine.com |accessdate=11 June 2019}}</ref>|-| 2018 || March || International expansion || DeepMind announces a new research lab in {{w|Paris}}, led by Remi Munos.<ref>{{cite web |title=A return to Paris |url=https://deepmind.com/blog/a-return-to-paris/ |website=deepmind.com |accessdate=5 July 2019}}</ref>
|-
| 2018 || May 9 || {{w|AI}} development || DeepMind develops a neural network loosely modeled on mammalian brains, that is better at navigating a maze than humans.<ref>{{cite web |title=DeepMind AI developed navigation neurons to solve a maze like us |url=https://www.newscientist.com/article/2168406-deepmind-ai-developed-navigation-neurons-to-solve-a-maze-like-us/ |website=newscientist.com |accessdate=1 June 2019}}</ref><ref>{{cite web |last1=Sample |first1=Ian |title=Google DeepMind's AI program learns human navigation skills |url=https://www.theguardian.com/technology/2018/may/09/googles-ai-program-deepmind-learns-human-navigation-skills |website=theguardian.com |accessdate=1 June 2019}}</ref><ref>{{cite web |last1=Gent |first1=Edd |title=This DeepMind AI Spontaneously Developed Digital Navigation ‘Neurons’ Like Ours |url=https://singularityhub.com/2018/05/14/this-deepmind-ai-spontaneously-developed-navigation-neurons-like-ours/ |website=singularityhub.com |accessdate=1 June 2019}}</ref><ref>{{cite web |last1=Quach |first1=Katyanna |title=DeepMind: Get a load of our rat-like AI. 'Ere, look. It solves mazes and stuff |url=https://www.theregister.co.uk/2018/05/10/deepmind_rat_like_ai_mazes/ |website=theregister.co.uk |accessdate=1 June 2019}}</ref>
|-
| 2018 || June 14 || {{w|AI}} development || DeepMind develops a neural network that teaches itself to ‘imagine’ a scene from different viewpoints, based on just a single image. The new type of computer vision algorithm can generate 3D models of a scene from 2D snapshots. It can tease out , unraveling details from the static images to guess at and solving spatial relationships, including the camera’s position. Dubbed a Generative Query Network (GQN), the system gets rid of labels and focuses on what's known as {{w|unsupervised learning}}.<ref>{{cite web |last1=Whyte |first1=Chelsea |title=DeepMind’s AI can ‘imagine’ a world based on a single picture |url=https://www.newscientist.com/article/2171675-deepminds-ai-can-imagine-a-world-based-on-a-single-picture/ |website=newscientist.com |accessdate=1 June 2019}}</ref><ref>{{cite web |last1=Bogle |first1=Ariel |title=Who needs humans? Google's DeepMind algorithm can teach itself to see |url=https://www.abc.net.au/news/science/2018-06-15/googles-deepmind-algorithm-can-teach-itself-to-see/9861590 |website=abc.net |accessdate=1 June 2019}}</ref><ref>{{cite web |last1=Wiggers |first1=Kyle |title=Google’s DeepMind develops AI that can render 3D objects from 2D pictures |url=https://venturebeat.com/2018/06/14/googles-deepmind-develops-ai-that-can-render-3d-objects-from-2d-pictures/ |website=venturebeat.com |accessdate=1 June 2019}}</ref><ref>{{cite web |title=Neural scene representation and rendering |url=https://science.sciencemag.org/content/360/6394/1204 |website=science.sciencemag.org |accessdate=1 June 2019}}</ref>
|-
| 2018 || June 15 || Controversy || The DeepMind Health Independent Reviewers’ 2018 report warns about the potential for DeepMind Health to be able to “exert excessive monopoly power” as a result of the data access and streaming infrastructure that’s bundled with provision of the Streams app, which would position DeepMind as the access-controlling intermediary between the structured health data and any other third parties.<ref>{{cite web |last1=Lomas |first1=Natasha |title=UK report warns DeepMind Health could gain ‘excessive monopoly power’ |url=https://techcrunch.com/2018/06/15/uk-report-warns-deepmind-health-could-gain-excessive-monopoly-power/ |website=techcrunch.com |accessdate=3 June 2019}}</ref><ref>{{cite web |last1=Lomas |first1=Natasha |title=Building health AIs should be UK ambition, says strategy review |url=https://techcrunch.com/2017/08/31/building-health-ais-should-be-uk-ambition-says-strategy-review/ |website=techcrunch.com |accessdate=3 June 2019}}</ref>
|-
| 2018 || June 26 || || {{w|Demis Hassabis}} is enlisted by the British government's {{w|Department for Digital, Culture, Media and Sport}} (DCMS), as an adviser to the Government’s new {{w|Office for Artificial Intelligence}}.<ref>{{cite web |title=DeepMind co-founder among experts to advise Government on using AI |url=https://www.eveningexpress.co.uk/news/deepmind-co-founder-among-experts-to-advise-government-on-using-ai/ |website=eveningexpress.co.uk |accessdate=3 June 2019}}</ref><ref>{{cite web |title=UK enlists DeepMind's Demis Hassabis to advise its new Government Office for AI |url=https://www.zdnet.com/article/uk-enlists-deepminds-demis-hassabis-to-advise-its-new-government-office-for-ai/ |website=zdnet.com |accessdate=3 June 2019}}</ref>
|-
| 2018 || November 19 || Partnership || DeepMind partners with {{w|OpenAI}} in a new paper that proposes a new method to train {{w|reinforcement learning}} agents in ways that enables them to surpass human performance. The paper, titled ''Reward learning from human preferences and demonstrations in Atari'', introduces a training model that combines human feedback and reward optimization to maximize the knowledge of RL agents.<ref>{{cite web |last1=Rodriguez |first1=Jesus |title=What’s New in Deep Learning Research: OpenAI and DeepMind Join Forces to Achieve Superhuman Performance in Reinforcement Learning |url=https://towardsdatascience.com/whats-new-in-deep-learning-research-openai-and-deepmind-join-forces-to-achieve-superhuman-48e7d1accf85 |website=towardsdatascience.com |accessdate=29 June 2019}}</ref>
|-
| 2018 || July || {{w|AI}} development || Researchers from DeepMind train one of its systems to play the famous computer game ''{{w|Quake III Arena}}''.<ref>[https://www.engadget.com/2018/07/03/deepmind-ai-quake-iii-arena-human/ "DeepMind AI’s new trick is playing ‘Quake III Arena’ like a human"]. ''Engadget''. 3 July 2018.</ref>
| 2018 || October 4 || {{w|AI}} development || DeepMind furthers cancer research and announces having been given access to {{w|mammogram}}s from roughly 30,000 women that were taken at {{w|Jikei University Hospital}} in {{w|Tokyo}}, {{w|Japan}} between 2007 and 2018. The data would be used to refine DeepMind's AI breast cancer detection algorithms.<ref>{{cite web |last1=Wiggers |first1=Kyle |title=DeepMind expands AI cancer research program to Japan |url=https://venturebeat.com/2018/10/04/deepmind-expands-ai-cancer-research-program-to-japan/ |website=venturebeat.com |accessdate=3 June 2019}}</ref><ref>{{cite web |last1=Shead |first1=Sam |title=Google DeepMind Given Access To Mammograms Of 30,000 Women By Japanese Hospital |url=https://www.forbes.com/sites/samshead/2018/10/04/google-deepmind-given-access-to-mammograms-of-30000-women-by-japanese-hospital/#280652659baa |website=forbes.com |accessdate=3 June 2019}}</ref>
|-
| 2018 || November 13 || Team || As part of a reorganization of its Google announces DeepMind’s health care efforts, DeepMind announces that its health division and the Streams app would unit to be absorbed into {{w|Google Health}}to the holding company to create an ‘AI assistant for nurses and doctors’.<ref>{{cite news |last=Vincent |first=James |url=https://www.theverge.com/2018/11/13/18091774/google-deepmind-health-absorbing-streams-team-ai-assistant-nurse-doctor|title=Google is absorbing DeepMind's health care unit to create an 'AI assistant for nurses and doctors' |work=The Verge |date=13 November 2018 |accessdate=28 May 2019}}</ref><ref>{{cite web |last1=Vincent |first1=James |title=Google is absorbing DeepMind’s health care unit to create an ‘AI assistant for nurses Is Absorbing DeepMind's Health Care Unit To Create An 'AI Assistant For Nurses and doctors’ Doctors' |url=https://wwwscience.thevergeslashdot.comorg/story/201818/11/13/180917742252225/google-deepmindis-absorbing-deepminds-health-absorbingcare-unit-to-streamscreate-teaman-ai-assistant-nursefor-nurses-and-doctors |website=science.slashdot.org |accessdate=11 June 2019}}</ref>|-| 2018 || November 19 || {{w|AI}} development || DeepMind safety research team publishes a paper on scalable agent alignment via reward modeling. The paper gives a summary of a research direction for solving the agent alignment problem.<ref>{{cite web |last1=Leike |first1=Jan |title=Scalable agent alignment via reward modeling |url=https://medium.com/@deepmindsafetyresearch/scalable-agent-alignment-via-reward-modeling-doctor bf4ab06dfd84 |website=thevergemedium.com |accessdate=30 May 11 June 2019}}</ref><ref>{{cite journal |last1=Leike |first1=Jan |last2=Krueger |first2=David |last3=Everitt |first3=Tom |last4=Martic |first4=Miljan |last5=Maini |first5=Vishal |last6=Legg |first6=Shane |url=https://arxiv.org/abs/1811.07871}}</ref>
|-
| 2018 || December 2 || {{w|AI}} development || DeepMind unveils AlphaFold, an {{w|algorithm }} able to predict the complex, three-dimensional shapes into which proteins can be folded. The prediction is based solely on their {{w|genetic sequence}}.<ref>{{cite web |last1=Lomas |first1=Natasha |title=DeepMind claims early progress in AI-based predictive protein modelling |url=https://techcrunch.com/2018/12/03/deepmind-claims-early-progress-in-ai-based-predictive-protein-modelling/ |website=techcrunch.com |accessdate=1 June 2019}}</ref><ref>{{cite web |last1=Wiggers |first1=Kyle |title=Deepmind’s AlphaFold wins CASP13 protein-folding competition |url=https://venturebeat.com/2018/12/03/deepminds-alphafold-wins-casp13-protein-folding-competition/ |website=venturebeat.com |accessdate=1 June 2019}}</ref><ref>{{cite web |last1=Kahn |first1=Jeremy |title=Alphabet's DeepMind AI Algorithm Wins Protein-Folding Contest |url=https://www.bloomberg.com/news/articles/2018-12-03/alphabet-s-deepmind-ai-algorithm-wins-protein-folding-contest |website=bloomberg.com |accessdate=1 June 2019}}</ref><ref>{{cite web |last1=Sample |first1=Ian |title=Google's DeepMind predicts 3D shapes of proteins |url=https://www.theguardian.com/science/2018/dec/02/google-deepminds-ai-program-alphafold-predicts-3d-shapes-of-proteins |website=theguardian.com |accessdate=1 June 2019}}</ref>
|-
| 2018 || December || Notable quote comment || {{w|Demis Hassabis}} announces: {{Quote|"I'd be much more pessimistic about the way the world is going to go if I didn't know there was something as game-changing as AI on the way."}}<ref name="Google DeepMind: the story behind the world's leading AI startup"/>
|-
| 2019 || February || Achievement || DeepMind's AI AlphaStar is revealed to outperform human professionals at {{w|StarCraft II}}, beating the humans 10 games in a row.<ref>{{cite web |last1=Grayson |first1=Nathan |title=Google's DeepMind AI Just Beat Two Pros At StarCraft II |url=https://kotaku.com/googles-deepmind-ai-just-beat-two-pros-at-starcraft-ii-1832034863 |website=kotaku.com |accessdate=3 June 2019}}</ref><ref>{{cite web |title=DeepMind Beats Pros at StarCraft in Another Triumph for Bots |url=https://www.wired.com/story/deepmind-beats-pros-starcraft-another-triumph-bots/ |website=wired.com |accessdate=3 June 2019}}</ref><ref>{{cite web |title=DeepMind AI AlphaStar goes 10-1 against top 'StarCraft II' pros |url=https://www.engadget.com/2019/01/24/deepmind-ai-starcraft-ii-demonstration-tlo-mana/ |website=engadget.com |accessdate=3 June 2019}}</ref><ref>{{cite web |title=DeepMind - StarCraft II Demonstration |url=https://starcraft2.com/en-us/news/22871520?linkId=100000004774117 |website=starcraft2.com |accessdate=28 May 2019}}</ref>
|-
| 2019 || February || {{w|AI}} development || DeepMind develops an algorithm aimed at boosting {{w|wind energy}} efficiency. Google reports having increased energy production by 20% after installing its own AI software across its largest renewable energy facilities in the United States.<ref>{{cite web |last1=Cuff |first1=Madeleine |title=Google and DeepMind deploy AI to predict wind energy output |url=https://www.businessgreen.com/bg/news/3071829/google-and-deep-mind-use-ai-to-predict-wind-energy |website=businessgreen.com |accessdate=4 June 2019}}</ref><ref>{{cite web |title=DeepMind: Google uses AI technology to boost wind energy efficiency |url=https://www.edie.net/news/8/Google-to-boost-wind-generation-efficiency-with-AI-powered-technology/ |website=edie.net |accessdate=4 June 2019}}</ref>
|-
| 2019 || March || Team || As of date, DeepMind has about 700 employees.<ref>{{cite book |last1=Webb |first1=Amy |title=The Big Nine: How the Tech Titans and Their Thinking Machines Could Warp Humanity |url=https://books.google.com.ar/books?id=-9RtDwAAQBAJ&pg=PT55&lpg=PT55&dq=deepmind+%22500..1500%22+employees&source=bl&ots=fxEkpTp_o-&sig=ACfU3U2g7oK7c2Kt5mQv7y6Pb-KJkJ_wgA&hl=en&sa=X&ved=2ahUKEwiw-6-Mu57jAhU_GbkGHQwsD28Q6AEwAHoECAkQAQ#v=onepage&q=deepmind%20%22500..1500%22%20employees&f=false}}</ref>
|-
| 2019 || April || Team || Google disbands the advisory board for DeepMind Health. It is the second disbanded review panel related to [[W:Alphabet Inc.|Alphabet]]'s AI dealings.<ref>{{cite web |last1=Fuertes |first1=Rechelle Ann |title=Google Disbands Advisory Board for DeepMind Health |url=https://edgy.app/google-deepmind-advisory-board |website=edgy.app |accessdate=3 June 2019}}</ref><ref>{{cite web |last1=Fuertes |first1=Rechelle Ann |title=Google Disbands Advisory Board for DeepMind Health |url=https://edgy.app/google-deepmind-advisory-board |website=edgy.app |accessdate=3 June 2019}}</ref><ref>{{cite web |last1=Merriman |first1=Chris |title=DeepMind becomes the second Alphabet company to disband an AI ethics panel |url=https://www.theinquirer.net/inquirer/news/3074256/google-deepmind-becomes-the-second-alphabet-company-to-close-an-ai-ethics-panel |website=theinquirer.net |accessdate=3 June 2019}}</ref>
|-
| 2019 || April || {{w|AI}} development || DeepMind researchers develop an {{w|AI}} tasked with teaching itself to solve arithmetic, algebra and probability problems, among others. However, the neural network performs poorly when tested on a maths exam taken by 16-year-olds in the United Kingdom, getting just 14 out of 40 questions correct, or the equivalent of an E grade.<ref>{{cite web |last1=Vaughan |first1=Adam |title=DeepMind created a maths AI that can add up to 6 but gets 7 wrong |url=https://www.newscientist.com/article/2198761-deepmind-created-a-maths-ai-that-can-add-up-to-6-but-gets-7-wrong/ |website=newscientist.com |accessdate=3 June 2019}}</ref><ref>{{cite web |last1=Srivastava |first1=Smriti |title=DeepMind Examined AI Neural Net Over High School Maths, but Lacked Success |url=https://www.analyticsinsight.net/deepmind-examined-ai-neural-net-over-high-school-maths-but-lacked-success/ |website=analyticsinsight.net |accessdate=3 June 2019}}</ref>
|-
| 2019 || June || {{w|AI}} development || DeepMind medical director Dominic King says {{w|AI}} could soon be used in predictive medicine, achieving this by looking at a medical record and electronic health record data to make predictions.<ref>{{cite web |title=DeepMind health lead says AI could soon be used in predictive medicine |url=https://www.digitalhealth.net/2019/06/deepmind-health-lead-says-ai-could-soon-be-used-in-predictive-medicine/ |website=digitalhealth.net |accessdate=29 June 2019}}</ref>
|-
|}
* [[Timeline of AlphaGo]]
* [[Timeline of machine learning]]
* [[Timeline of OpenAI]]
==External links==
62,430
edits

Navigation menu