Last year the annual artificial intelligence conference NeurIPS invited 230 researchers from Africa to attend a "Black in AI" workshop in Montreal.It was a great opportunity to bring some diversity to the field of AI.The Partnership on AI—a group founded by Amazon, Facebook, Google's DeepMind subsidiary, Microsoft, and IBM—contends that these sorts of visa issues are a threat to the development of AI.To keep ideas moving across borders, the report argues that governments should create special classes of visa to enable "global freedom of movement" for AI experts.This would include not just short term visas for academic conferences, but faster and more generous visa processing for AI professionals looking to work in a country long-term.The Trump administration, which already tightened the limits on the number of refugees allowed into the country, is considering deep cuts to the US refugee program that could all but destroy it.
Between the fake news potential of deepfakes, the fear of robots stealing jobs, and the occasional call for automated systems to have control of the nuclear button, A.I.’s public image could do with a PR makeover here in 2019.Combining genomics, big data analysis, and deep learning, the company — which is based in Rockville in Johns Hopkins University’s Emerging Technology Centers — has been using artificial intelligence algorithms to potentially discover the next world-changing drug.techniques of the moment, it’s found a way of discovering drug molecules not only far more cheaply than usual, but also much, much, much faster.“There is a new concept in artificial intelligence called Generative Adversarial Networks (GANs), which was first introduced in 2014,” Alex Zhavoronkov, CEO of Insilico, told Digital Trends.The worst application seen to date were the deepfakes.”A Generative Adversarial Network consists of two distinct neural networks: a generator and a discriminator.
A study by Hong Kong-based Insilico Medicine utilized a new AI system that created a series of novel molecules capable of treating fibrosis and other illnesses on a significantly shorter timeline, the South China Morning Post reported.Why it matters: Insilico used AI to identify treatment candidates in just three weeks compared with the traditional drug discovery process, which can take decades and cost billions.Besides saving money for pharmaceutical companies, streamlining the development of new drugs could also save lives.Details: Out of the six molecules that the firm successfully created, one was later found to be effective in treating mice with renal fibrosis.Insilico’s system dubbed GENTRL is powered by a “generative chemistry that utilizes modern AI techniques” similar to Google’s Go-playing DeepMind, according to BioSpace.GENTRL was developed in collaboration with pharmaceutical services contractor Wuxi Apptec and the University of Toronto’s Alán Aspuru-Guzik.
Alex Zhavoronkov, CEO of Insilico Medicine, a startup that generates potential drugs using artificial intelligence, was recently given a challenge by one of his pharma company partners.His team would see how quickly Insilico’s AI could identify new molecules that bind with a protein associated with tissue scarring.There’s hope—and hype—that AI could help chip away at that figure by reducing the time and labor before a drug starts clinical trials.The idea is that the same techniques used to generate realistic deepfakes and deftly play Go might be able to decipher the complex rules of drug design and generate molecules from scratch.There are signs AI has potential.In December, Alphabet’s DeepMind debuted AlphaFold, an algorithm designed to predict protein folding—an important step for identifying potential disease targets.
Viewing scenes and making sense of them is something people do effortlessly every day.That’s patently untrue of most AI systems, which tend to reason rather poorly.But emerging techniques in visual recognition, language understanding, and symbolic program execution promise to imbue them with the ability to generalize to new examples, much like humans.Scientists at the MIT-IBM Watson AI Lab, a joint 10-year, $240 million partnership to propel scientific breakthroughs in machine learning, are perfecting an approach they say might overcome longstanding barriers in AI model design.It marries deep learning with symbolist philosophies, which advocate representations and logical rules as intelligent machine cornerstones, to create programs that learn about the world through observation.“You have a visual perception challenge — you have a question and you have to understand what those words mean — and then you have a logic reasoning part that you have to execute to solve this problem [as well].”
Tech giants are training artificial intelligence (AI) computers to be game masters as China strives for world AI leadership within the next ten years, according to leaders speaking at the World Artificial Intelligence Conference (WAIC) in Shanghai on Thursday.Broadening AI research to Artificial General Intelligence (AGI), where a machine is trained to perform any intellectual task that a human is capable of, is being accelerated in China, and the modeling and simulation in virtual reality is a crucial step for the great leap, Tencent CEO Pony Ma said at the conference.Earlier this month, Tencent Wukong AI, an autonomous system devised by the company, faced off with a professional human team playing the company’s hugely popular game, Honour of Kings, in an international competition in Malaysia.The Chinese gaming giant developed its own computer for the complex board game Go, Fine Art—like Alphabet’s DeepMind research project AlphaGo—in early 2016, which later beat China’s top professional player, Ke Jie, in January last year.Harry Shum, the company’s executive vice president, said the company had made “the world’s best AI system in the field of mahjong,” which earned top ranking, the 10th dan, on international professional mahjong platform Tianfeng in June, a level that fewer than 20 humans have reached.The US tech giant’s Mahjong AI Suphx, developed by Microsoft Asia Research Institute, surpassed the average score from 10th dan-ranked human players after playing more than 5,000 games on the platform beginning in March.
Baidu, the most ardent AI advocate in China, may have lost a lot more.Worse yet, its wager on AI may go the way of its failed online-to-offline push from several years ago.There is no question that companies must invest in AI to succeed in the future, and losses are natural in the initial stages of developing and commercializing such an emerging technology.The failure or success of its AI strategy will play a more critical role in determining the future of Baidu.Worse yet, the possibility of a handsome return appears dim.Baidu’s Apollo autonomous driving unit is its most capital-intensive.
In his 2017 Amazon shareholder letter, Jeff Bezos wrote something interesting about Alexa, Amazon’s voice-driven intelligent assistant:In the U.S., U.K., and Germany, we’ve improved Alexa’s spoken language understanding by more than 25% over the last 12 months through enhancements in Alexa’s machine learning components and the use of semi-supervised learning techniques.Given those results, it might be interesting to try semi-supervised learning on our own classification problems.Supervised learning starts with training data that are tagged with the correct answers (target values).In general, tagging data costs money and takes time.Semi-supervised learning goes back at least 15 years, possibly more; Jerry Zhu of the University of Wisconsin wrote a literature survey in 2005.
Back in February, Google announced a series of updates to its Google Cloud Platform (GCP) AI text-to-speech and speech-to-text services that introduced multichannel recognition, device profiles, and additional languages synthesized by an AI system — WaveNet — pioneered by Google parent company Alphabet’s DeepMind.Building on those enhancements, the Mountain View company today expanded the number of new variants and voices in Cloud Text-to-Speech by nearly 70%, boosting the total number of languages and variants covered to 33.Now, thanks to the addition of 76 new voices and 38 new WaveNet-powered voices, Cloud Text-to-Speech boasts 187 total voices (up from 106 at the beginning of this year) and 95 total WaveNet voices (up from 57 in February and 6 a year and a half ago).Among the newly supported languages and variants are Czech, English (India), Filipino, Finnish, Greek, Hindi, Hungarian, Indonesian, Mandarin Chinese (China), Modern Standard Arabic, Norwegian (Nynorsk), and Vietnamese, all of which have at least one AI-generated voice.“With these updates, Cloud Text-to-Speech developers can now reach millions more people across numerous countries with their applications — with many more languages to come,” wrote product manager Dan Aharon.“This enables a broad range of use cases, including call center IVR, interacting with IoT devices in cars and the home, and audio-enablement of books and other text-based content.”
Reinforcement learning, the AI training technique that’s brought to fruition systems capable of defeating world poker champions and guiding self-driving cars, isn’t the simplest thing in the world to wrangle.That’s particularly true in the gaming domain, where cutting-edge approaches sometimes require bespoke tools that aren’t publicly available.In a paper recently published on the preprint server Arxiv.org, researchers at Alphabet’s DeepMind describe a game-oriented reinforcement learning framework dubbed OpenSpiel.At its core, it’s a collection of environments and algorithms for research in general reinforcement learning and search and planning in games, with tools to analyze learning dynamics and other common evaluation metrics.“The purpose of OpenSpiel is to promote general multiagent reinforcement learning across many different game types, in a similar way as general game-playing but with a heavy emphasis on learning and not in competition form,” wrote the researchers.“We hope that OpenSpiel could have a similar effect on general [reinforcement learning] in games as the Atari Learning Environment has had on single-agent [reinforcement learning].”
The UK’s health data watchdog, the National Data Guardian (NDG), has published correspondence between her office and the national privacy watchdog which informed the ICO’s finding in 2017 that a data-sharing arrangement between an NHS Trust and Google-owned DeepMind broke the law.In fall 2015 the Royal Free NHS Trust and DeepMind signed a data-sharing agreement which saw the medical records of 1.6 million people quietly passed to the AI company without patients being asked for their consent.DeepMind has also since announced it plans to transfer its health division to Google.Although — to our knowledge — no NHS trusts have yet signed new contracts for Streams with the ad giant.In a subsequent audit of Streams that was a required by the regulator, the trust’s law firm, Linklaters, argued that a call on whether a duty of confidentiality has been breached should be judged from the point of view of the clinician’s conscience, rather than the patient’s reasonable expectations.“I do champion innovative technologies and new treatments that are powered by data.
Google-owned AI company DeepMind said on Wednesday that cofounder Mustafa Suleyman was going on leave.Suleyman tweeted on Thursday to quell speculation that had started to swirl following news of his departure.He said he's taking some "personal time for a break to recharge."Visit Business Insider's homepage for more stories.DeepMind cofounder Mustafa Suleyman put out a tweet addressing the questions surrounding his temporary departure from the company."Mustafa's taking some time out right now after ten hectic years," a spokeswoman said.She emphasised that the decision was mutual, and that he was expected to return at the end of the year.
The co-founder of Google’s London-based DeepMind artificial intelligence unit has taken an extended leave of absence from the firm.According the FT, co-founder Mustafa Suleyman is currently on extended leave from the company, amid speculation that parent company Google had taken over the bulk of his responsibilities.The development comes after Google announced last November that it would transfer control of DeepMind to a new Google Health division in California, as it invests more in commercialising its medical research efforts.That said, DeepMind was reportedly intending to retain and expand its London base.According the FT, DeepMind said that Suleyman’s leave, was a mutual decision between him and Alphabet.“Mustafa’s taking some time out right now after ten hectic years,” a DeepMind spokesperson is quoted as saying.
After vanquishing the best humanity has to offer in the ancient game of Go, also known as weiqi, Google is now looking to the massively popular game of football to train its next wave of AI technology to “bend it like Beckham.”The US internet giant published research in June revealing that its “Brain Team” is working on a game known as Google Research Football Environment to train smart agents that can interact with their environment to solve complex tasks, providing insights into real-world AI applications such as autonomous driving and robotics.Google released a beta version of Football Environment as open-source code on Github earlier this year.The game was built using a publicly available title called Gameplay Football and uses advanced game simulation, including goals, fouls, corners, penalty kicks, and offside plays, according to the announcement on Google’s AI blog.The move comes as technology giants push the boundaries of AI technology, a form of machine learning that has been dubbed the Fourth Industrial Revolution, as it moves into more corners of everyday life, from autonomous driving, smart city infrastructure, and internet-of-things applications to workplace automation.Remove this ad space by subscribing.