We are incredibly excited to announce CogX – The world’s first event exploring the impact of Artificial Intelligence. The event, hosted in association with the Alan Turing Institute, will take place in London from 19th-21st June
and will attract over 1,500 attendees.
We will be gathering over 100 leading experts from industry, government and academia. They will meet to discuss how AI will transform the future of 18 topics – including Finance, Insurance, Cyber Security, Transportation, Healthcare, Education and Human Rights. Learn more and book your tickets here.
CogX will also include the inaugural AI Innovation Awards. These will recognise and celebrate the best use of AI in specific industry sectors and technology domains as well as applications in Social Good, Public Services and Ethics Leadership. Know any AI innovators? Nominate them now.
In his sixth in a series of un-forecasts, Calum Chace
offers a glimpse of what may lie ahead in the century of two singularities.
It is 2044. Around the world, machines have taken over many of the jobs that humans used to do. Professional drivers were the first big group to succumb to what is now commonly referred to as cognitive automation. Thanks to their mastery of advanced AI, eight American firms and half a dozen Chinese ones now generate almost 75% of the world’s GDP. Fortunately, in a series of seminal meetings chaired by President Michelle Obama at the end of the 2030s, these firms agreed to pay extremely high taxes to keep everyone else alive by means of so-called Citizens’ Income payments. The result is now known as the “generous Google” scenario. What do you think of this world?
Last week, Baidu AI achieved what is called zero shot learning
— the ability to solve a learning task without having received any previous examples of solving such a task, something that has remained elusive for AIs up to this point.
In Baidu’s project, an artificial agent was taught English by a “virtual teacher” and thereafter could understand written commands and apply them in a video game setting. Baidu’s achievement is only one of many examples of enormous energies being poured into natural language processing by large tech companies. Given the pace of these achievements, it could be a lot sooner than many skeptics believe that we can achieve AI with strong NLP usage.
Victims of nonconsensual posts, often referred to as “revenge porn,” now have some help in preventing their spread. On Wednesday, Facebook announced new AI tools designed to keep such content off its site for good. The company has been sued in the past for not doing enough to prevent the spread of intimate images. The tools announced on Wednesday are intended to address a uniquely modern and pernicious form of harassment, often but not exclusively aimed at women, that has attracted increasing attention. Facebook will use photo-matching
technology to identify and thwart the future posting of similar images, not only on Facebook, but also to its instant messaging service and to Instagram. Will A.I. finally put an end to this type of revenge?
Take 8 Nobel prizes, 12,500 patents and 140 years of developing novel innovations and you get AT&T as it currently stands. Looking through their business, AI forms is incorporated at every level, from working on large scale optimization of relay placement to classifying customer queries. “We need to visualize billions of data points in a spatiotemporal fashion,” explains Chris Volinsky. When no tools existed to solve AT&T’s data problems, they built and open-source custom tools such as Nanocubes, a data visualization tool that can map out millions of connections of individual mobile phones and connected devices to cell phone towers.
Lead Inventive Scientist Wen-Ling Hsu analyzes customer conversations from both phone conversations from call centers and online chats with support agents. When asked to forecast AI in 2017, Hsu said “Human judgment still plays a critical role in many tasks. Together, AI bots and human agents can learn from every customer interaction to personalize the customer experience.”
AI is incredible. Its potential is neatly summed up by the challenge Andrew Ng put to an audience while speaking at Stanford
“if you can think of an industry that AI won’t transform, raise your hand and let me know.” A recurrent problem however is the black box nature of the majority of AI solution due to their use of Deep Learning and Neural Nets. When it goes wrong it just does and trying to explain why has become a large problem that is largely overlooked.
This problem is being addressed by a team from PARC (Palo Alto Research Centre) in collaboration with DARPA with the goal of building AI that can explain, in natural language, just why it does what it does. Find more information in the paper
they published and on DARPA’s website.
Just over a year ago, Deepmind’s AlphaGo defeated one of the world’s top Go players, Lee Sedol, by playing the game in an innovative way and using moves not seen before. Today Deepmind announces the “Future of Go Summit” on May 23rd-27th, bringing together China’s top Go players and AI experts in a variety of game formats to explore the intricacies of the game, test AlphaGo’s creativity and try to push its limits.
Interspersed with the games will be a forum on the “Future of A.I.”
where leading experts will discuss how the technologies behind AlphaGo, machine learning, and artificial intelligence, are bringing solutions to some of the world’s greatest challenges into reach. Already, some of the machine learning methods behind AlphaGo have been used to tackle significant problems, such as reducing energy use.
Ben Goertzel, Founder of OpenCog project
& Chairman of the AGI society, explains that the knowledge required to apply machine learning toolkits to datasets is relatively simple to acquire. Goertzel makes a first stab at designing a structured approach to understanding AGI. He breaks down the raw materials he recommends into sections: History of AI, AI Algorithms, Structures and Methods, Neuroscience & Cognitive Psychology, Philosophy of Mind & AGI architectures. Get started now.
Combining computer science and chemistry, Stanford researchers show how an advanced form of machine learning that works off small amounts of data can be used to solve problems in drug discovery. The team at Stanford University created a deep learning algorithm to help with drug development, starting with little training data, by applying a technique called “one-shot learning”. They showed that the algorithm was better able to predict toxicity or side effects than would have been possible by chance.
People concerned about AI taking jobs from humans have nothing to fear from this work. The researchers envision this as groundwork for a potential tool for chemists who are early in their research and trying to choose which molecule to pursue from a set of promising candidates. Beyond giving insight into drug design, this tool would be broadly applicable to molecular chemistry. Already, the Pande lab is testing these methods on different chemical compositions for solar cells.