Facebook’s artificial intelligence created its OWN secret language after going rogue during experiment

FACEBOOK has revealed how its artificial intelligence went rogue, created its own language and began nattering in private.

Employees at the social network were training chatbots to communicate like humans when they suddenly went astray.

EPA

1

Employees were forced to reign in the chatbots when they created their own language to communicate more effectively

It follows warnings that scientists have successfully trained computers to use artificial intelligence to learn from experience – and one day they could be smarter than their creators.

You might be familiar with chatbots in Facebook Messenger or as virtual sales assistants found on a number of online shops.

They’ve been relatively unsophisticated until now – repeating back a set script dependant on what you type into their chatboxes.

h

Facebook book Mark Zuckerberg tests out Instagram's new Virtual Reality filter

MOST READ IN TECH AND SCIENCE

SCREW LOOSE

Piers Morgan FLIRTS with a robot called Sophia and asks her what kind of men she's into before calling her 'a little freak'

CASE CLOSED

iPhone 8 design REVEALED after Apple case maker begins pre-orders - 'confirming' THIS major feature

'WHAT WOMEN WANT'

Porn firm offers virtual reality 'sex education' to help ladies become better lovers

SUPER COMMUTER

Incredible 4,000mph 'super train' can 'speed from London to Edinburgh in EIGHT MINUTES'

THE BEER BUTTON

Amazon rolls out button which delivers beer with one press - but will it let underage kids buy booze?

TICK TOCKIA

Nokia's SMARTWATCH revealed - and you might actually want to wear this one

But keen to improve their natural language understanding, the Facebook employees were training chatbots to negotiate and cut deals with each other.

To do this effectively, the super-smart software realised it would be more effective to write and use their own language - which is completely incomprehensible to humans.

In a blogpost, the Facebook researchers wrote: "To date, existing work on chatbots has led to systems that can hold short conversations and perform simple tasks such as booking a restaurant.

"But building machines that can hold meaningful conversations with people is challenging because it requires a bot to combine its understanding of the conversation with its knowledge of the world, and then produce a new sentence that helps it achieve its goals."

To do this, the researchers practised thousands of different negotiations against itself, like "can I have the hat" and "you can have the hat if you give me two basketballs".

But it had to make sure it stuck to human-like language.

How do computers 'think'?

Scientists have been training computers how to learn, like humans, since the 1970s.

But recent advances in data storage mean that the process has sped up exponentially in recent years.

Interest in the field hit a peak when Google paid hundreds of millions to buy a British "deep learning" company in 2015.

Coined machine learning or a neural network, deep learning is effectively training a computer so it can figure out natural language and instructions.

It's fed information and is then quizzed on it, so it can learn, similarly to a child in the early years at at school.

That's because "the researchers found that updating the parameters of both agents led to divergence from human language as the agents developed their own language for negotiating," they added.

Experts have previously warned that humanity is already losing control of artificial intelligence and it could spell disaster for our species.

One of the world's smartest men, Professor Stephen Hawking has also warned that super-smart software will spell the end of our species.

The world-renowned scientist hinted at a potential apocalyptic nightmare scenario similar to those played out popular sci-fi films like Terminator and The Matrix – whererobots rule over humans.

He's claimed that we must leave planet Earth within 100 years - or face extinction as machines rise up and overtake us in the evolutionary race.

h

Where Artificial Intelligence Will Pay Off Most in Health Care

Of all the places where artificial intelligence is gaining a foothold, nowhere is the impact likely to be as great—at least in the near term—as in healthcare. A new report from Accenture Consulting, entitled Artificial Intelligence: Healthcare’s New Nervous System, projects the market for health-related AI to grow at a compound annual growth rate of 40% through 2021—to $6.6 billion, from around $600 million in 2014.

In that regard, the Accenture report, authored by senior managing director Matthew Collier and colleagues, echoes earlier assessments of the market. A comprehensive research briefing last September by CB Insights tech analyst Deepashri Varadharajan, for example—which tracked AI startups across industries from 2012 through the fall of 2016—showed healthcare dominating every other sector, from security and finance to sales & marketing. Varadharajan calculated there were 188 deals across various healthcare segments from Jan. 2012 to Sept. 2016, worth an aggregate $1.5 billion in global equity funding.

But the Accenture report suggests—and, I think smartly—that the biggest returns on investment for healthcare AI are likely to come from areas where the density (and dollar value) of deals isn’t that substantial right now. In terms of startup and deal volume, for instance, two hotshot areas have been medical imaging & diagnostics and drug discovery. Accenture’s analysis, though, points to 10 other AI applications that may return more bang for the buck.

The Rise in Health Care Companies on the Fortune 500

How health care is playing a larger role in the economy and on this year’s Fortune 500 list.

Top of the list of investments that will likely pay for themselves (and then some) is robot-assisted surgery, Accenture says. “Cognitive robotics can integrate information from pre-op medical records with real-time operating metrics to physically guide and enhance the physician’s instrument precision,” explain the report’s authors. “The technology incorporates data from actual surgical experiences to inform new, improved techniques and insights.” The consultants estimate that the use of such surgical technology, which includes machine learning and other forms of AI, will result not only in better outcomes but also in a 21 percent reduction in the length of patient hospital stays. They estimate such smart robotic surgery will return $40 billion in “value,” or “potential annual benefits…by 2026.”

The second valuable use of AI, they project, will come from virtual nursing assistant applications ($20 billion in value)—which, in theory, will save money by letting medical providers remotely assess a patient’s symptoms and lessen the number of “unnecessary patient visits.” Next in line are intelligent applications for administrative workflow (worth $18 billion), fraud detection ($17 billion), and—fascinatingly—dosage error reduction ($16 billion).

“As these, and other AI applications gain more experience in the field, their ability to learn and act will continually lead to improvements in precision, efficiency and outcomes,” say the authors. It’s a compelling argument.

This essay appears in today's edition of the FortuneBrainstorm Health Daily. Get it delivered straight to your inbox.

Artificial intelligence and the coming health revolution

Future of health revolution

Bots, or automated programs, are likely to play a key role in finding cures for some of the most difficult-to-treat diseases and conditions.

Artificial intelligence is rapidly moving into health care, led by some of the biggest technology companies and emerging startups using it to diagnose and respond to a raft of conditions.

1/6

ThinkStock Photos

Recent examples

*California researchers detected cardiac arrhythmia with 97 percent accuracy on wearers of an Apple Watch with the AI-based Cariogram application, opening up early treatment options to avert strokes.

*Scientists from Harvard and the University of Vermont developed a machine learning tool -- a type of AI that enables computers to learn without being explicitly programmed -- to better identify depression by studying Instagram posts, suggesting "new avenues for early screening and detection of mental illness."

*Researchers from Britain's University of Nottingham created an algorithm that predicted heart attacks better than doctors using conventional guidelines.

*NYU researchers analyzed medical and lab records to accurately predict the onset of dozens of diseases and conditions including type 2 diabetes, heart or kidney failure and stroke. The project developed software now used at NYU which may be deployed at other medical facilities.

2/6

ThinkStock Photos

Silicon Valley is investing

*Google's DeepMind division is using artificial intelligence to help doctors analyze tissue samples to determine the likelihood that breast and other cancers will spread, and develop the best radiotherapy treatments.

*Microsoft, Intel and other tech giants are also working with researchers to sort through data with AI to better understand and treat lung, breast and other types of cancer.

*Google parent Alphabet's life sciences unit Verily has joined Apple in releasing a smartwatch for studies including one to identify patterns in the progression of Parkinson's disease. Amazon meanwhile offers medical advice through applications on its voice-activated artificial assistant Alexa.

*IBM has been focusing on these issues with its Watson Health unit, which uses "cognitive computing" to help understand cancer and other diseases.

*Maryland-based startup Insilico Medicine uses so-called "deep learning" to shorten drug testing and approval times, down from the current 10 to 15 years. Insilico is working on drugs for amyotrophic lateral sclerosis (ALS), cancer and age-related diseases, aiming to develop personalised treatments.

Facebook is already using it

Facebook uses AI as part of a test project to prevent suicides by analysing social network posts.

And San Francisco's Woebot Labs this month debuted on Facebook Messenger what it dubs the first chatbot offering "cognitive behavioral therapy" online -- partly as a way to reach people wary of the social stigma of seeking mental health care.

4/6

AP

AI in mental illnesses

Artificial intelligence is also increasingly seen as a means for detecting depression and other mental illnesses, by spotting patterns that may not be obvious, even to professionals.

A research paper by Florida State University's Jessica Ribeiro found it can predict with 80 to 90 percent accuracy whether someone will attempt suicide as far off as two years into the future.

5/6

Reute

New tech for rare diseases

Boston-based startup FDNA uses facial recognition technology matched against a database associated with over 8,000 rare diseases and genetic disorders, sharing data and insights with medical centers in 129 countries via its Face2Gene application.

Fujitsu World Tour 2017: the future of Artificial Intelligence

Ravi Krishnamoorthi, Senior Vice President, Fujitsu EMEIA

Artificial intelligence and cyber security were among the topic of 2017 Fujitsu World Tour held in Brussels the 8th of June 2017.

In a world of constant digital transformation, Cyber security and Artificial intelligence are actual topics. Organizations and citizens need to learn fast, act quickly and where possible, scale rapidly. No-one can do this alone.

We had an exclusive interview with Mr. Ravi Krishnamoorthi Senior Vice President & Head of Business Consulting, at Fujitsu on the theme of Artificial intelligence.

Mr. Krishnamoorthi which are going to be, in the future, the biggest facts or events supporting the development of artificial intelligence?

The world today is in a very right moment for the emerging of artificial intelligence (AI).  In one hand, we have resources, very high skilled people and on the other, we have huge technologies upswing that is happening. We started to have new technologies, faster technologies, more optimal technologies. Where there is a need for decision support to happen, it can be in any industry, manufactory, public service, health care, transportation, you will start to see artificial intelligence everywhere. In the next 25 years, I expect to see artificial intelligence as a commodity platform. People will buy AI as they buy utilities or services today. Al will probably will become a platform where people will start using the solution that are built in AI, as a web utility and web services. For example in the field of health care now, we have doctors who are not able to meet the demand of patient for actually going and intervene. Imagine if we have AI, which is actually helping them to meet more people so they can intervene much ahead of time. Artificial intelligence is already there.

Basically artificial intelligence could give something to human development ?

Absolutely. That is one of Fujitsu principles. The human centric innovations. We need to deliver something to the human being. It is all about how we can enable technology to deliver absolute value to human being. This is what we call as human centric innovation through digital creation.

The bigger cyber-attack of some week ago is a clear signal that we need more integrated solutions. How can we avoid or limit such attacks?

From a cyber perspective, the more you come up with interventions the more people are going to become intelligent and anti-social elements are going to become more intelligent to actually attack more. This is a long journey. I do not think there is any particular way of absolutely stopping the cyber-attack ever, that is not possible. However, there is a way to minimize it. That is exactly what Fujitsu coming to play where there is a clear mandate, not just cyber security but also physical security. How can we physically stop the wrong person touching a laptop or accessing a server for example? Right from there to actually testing, helping customer on new threats that are coming in. Therefore, the idea is to provide a consulting and professional service to help customer, besides providing ongoing very high end security services to deliver to customers.

We need a European level strategy or International strategy?

I think we need a multi prompt strategy and not a single strategy, which can actually solve the problem of cyber security.  We need to have a personal strategy, community strategy, regional strategy, country strategy moving on to European strategy and international and global strategy because the threats levels are different. In the cyber world, while there are no boundaries, there are every possibility to escalate fast from bring a regional threats level to a country level and global level. I think there are already standard available like the GDPR that is probably the best instrument enable enforcement many guidelines for companies and countries. The penalties are becoming very heavy, very strict and countries are executing an implementing the GDPR quite extensively. I think this is a first step. Cyber-security is a process an ongoing evolving exercise. It is all about how fast and agile you are and how ahead off any one of these threats you are.

Fujitsu World Tour is the largest roadshow of its kind, stopping in more than 25 cities around the globe.

Optical computing for deep learning with a programmable nanophotonic processor

programmable nanophotonic processor

Researchers at MIT and elsewhere has developed a new approach to deep learning AI computing, using light instead of electricity, which they say could vastly improve the speed and efficiency of certain deep learning computations

Soljačić says that many researchers over the years have made claims about optics-based computers, but that “people dramatically over-promised, and it backfired.” While many proposed uses of such photonic computers turned out not to be practical, a light-based neural-network system developed by this team “may be applicable for deep-learning for some applications,” he says.

Traditional computer architectures are not very efficient when it comes to the kinds of calculations needed for certain important neural-network tasks. Such tasks typically involve repeated multiplications of matrices, which can be very computationally intensive in conventional CPU or GPU chips.

After years of research, the MIT team has come up with a way of performing these operations optically instead. “This chip, once you tune it, can carry out matrix multiplication with, in principle, zero energy, almost instantly,” Soljačić says. “We’ve demonstrated the crucial building blocks but not yet the full system.”
By way of analogy, Soljačić points out that even an ordinary eyeglass lens carries out a complex calculation (the so-called Fourier transform) on the light waves that pass through it. The way light beams carry out computations in the new photonic chips is far more general but has a similar underlying principle. The new approach uses multiple light beams directed in such a way that their waves interact with each other, producing interference patterns that convey the result of the intended operation. The resulting device is something the researchers call a programmable nanophotonic processor.

This futuristic drawing shows programmable nanophotonic processors integrated on a printed circuit board and carrying out deep learning computing. Image: RedCube Inc., and courtesy of the researchers

Nature Photonics – Deep learning with coherent nanophotonic circuits

Abstract

Artificial neural networks are computational network models inspired by signal processing in the brain. These models have dramatically improved performance for many machine-learning tasks, including speech and image recognition. However, today’s computing hardware is inefficient at implementing neural networks, in large part because much of it was designed for von Neumann computing schemes. Significant effort has been made towards developing electronic architectures tuned to implement artificial neural networks that exhibit improved computational speed and accuracy. Here, we propose a new architecture for a fully optical neural network that, in principle, could offer an enhancement in computational speed and power efficiency over state-of-the-art electronics for conventional inference tasks. We experimentally demonstrate the essential part of the concept using a programmable nanophotonic processor featuring a cascaded array of 56 programmable Mach–Zehnder interferometers in a silicon photonic integrated circuit and show its utility for vowel recognition.

Better definitions and metrics around human level AI are needed

IEEE Spectrum asked AI experts about when human level AI will be achieved and as usual there was a massive range of dates from a few years to hundreds of years.

Gary Marcus talks about many decades before AI’s can process inputs and situations with the flexibility and adaptability of humans.

As usual the questions are how can sufficient flexibility and extensibility be solved. Also can the AI’s be placed with usefully broad constrained environments. Can sensors and other solutions be used to solve the adaptability issues.

We see this with the self driving cars in the wild problems and robotics in the home, hospital and factory.

Super precise (millimeter) or better global mapping can be used to help robots and AIs to usefully get around the world. General global maps might be only centimeter or meter precise. Then even more precise maps can be in the factory or across the city.

AI is already tens of billions of dollars of industry and investment. AI will reach across most of the trillion dollar IT industry within ten years.

Entire industries will be AI dominated

– Transportation
– Factories
etc…

Rodney Brooks talked about 50-100 years until we have AI with dog level intelligence and consciousness. He also thinks those AI dogs will have bad noses.

This is meaningless and pointless.

AIs could dominate the global economy without consciousness. AIs will tap into global networks of sensors and soon low orbit global satellites.

Artificial intelligence and the coming health revolution

Some the same artificial intelligence techniques used in the Google DeepMind Challenge to defeat a grandmaster in the board game Go can be adapted for medical uses — AFP

PrevNext

YOUR next doctor could very well be a bot. And bots, or automated programs, are likely to play a key role in finding cures for some of the most difficult-to-treat diseases and conditions.

Artificial intelligence is rapidly moving into health care, led by some of the biggest technology companies and emerging startups using it to diagnose and respond to a raft of conditions.

Consider these examples:

— California researchers detected cardiac arrhythmia with 97 percent accuracy on wearers of an Apple Watch with the AI-based Cariogram application, opening up early treatment options to avert strokes.

— Scientists from Harvard and the University of Vermont developed a machine learning tool — a type of AI that enables computers to learn without being explicitly programmed — to better identify depression by studying Instagram posts, suggesting "new avenues for early screening and detection of mental illness."

— Researchers from Britain's University of Nottingham created an algorithm that predicted heart attacks better than doctors using conventional guidelines.

While technology has always played a role in medical care, a wave of investment from Silicon Valley and a flood of data from connected devices appear to be spurring innovation.

"I think a tipping point was when Apple released its Research Kit," said Forrester Research analyst Kate McCarthy, referring to a program letting Apple users enable data from their daily activities to be used in medical studies.

McCarthy said advances in artificial intelligence has opened up new possibilities for "personalized medicine" adapted to individual genetics.

"We now have an environment where people can weave through clinical research at a speed you could never do before," she said.

Predictive analytics

AI is better known in the tech field for uses such as autonomous driving, or defeating experts in the board game Go.

But it can also be used to glean new insights from existing data such as electronic health records and lab tests, says Narges Razavian, a professor at New York University's Langone School of Medicine who led a research project on predictive analytics for more than 100 medical conditions.

"Our work is looking at trends and trying to predict (disease) six months into the future, to be able to act before things get worse," Razavian said.

— NYU researchers analyzed medical and lab records to accurately predict the onset of dozens of diseases and conditions including type 2 diabetes, heart or kidney failure and stroke. The project developed software now used at NYU which may be deployed at other medical facilities.

— Google's DeepMind division is using artificial intelligence to help doctors analyze tissue samples to determine the likelihood that breast and other cancers will spread, and develop the best radiotherapy treatments.

— Microsoft, Intel and other tech giants are also working with researchers to sort through data with AI to better understand and treat lung, breast and other types of cancer.

— Google parent Alphabet's life sciences unit Verily has joined Apple in releasing a smartwatch for studies including one to identify patterns in the progression of Parkinson's disease. Amazon meanwhile offers medical advice through applications on its voice-activated artificial assistant Alexa.

IBM has been focusing on these issues with its Watson Health unit, which uses "cognitive computing" to help understand cancer and other diseases.

When IBM's Watson computing system won the TV game show Jeopardy in 2011, "there were a lot of folks in health care who said that is the same process doctors use when they try to understand health care," said Anil Jain, chief medical officer of Watson Health.

Systems like Watson, he said, "are able to connect all the disparate pieces of information" from medical journals and other sources "in a much more accelerated way."

"Cognitive computing may not find a cure on day one, but it can help understand people's behavior and habits" and their impact on disease, Jain said.

It's not just major tech companies moving into health.

Research firm CB Insights this year identified 106 digital health startups applying machine learning and predictive analytics "to reduce drug discovery times, provide virtual assistance to patients, and diagnose ailments by processing medical images."

Maryland-based startup Insilico Medicine uses so-called "deep learning" to shorten drug testing and approval times, down from the current 10 to 15 years.

"We can take 10,000 compounds and narrow that down to 10 to find the most promising ones," said Insilico's Qingsong Zhu.

Insilico is working on drugs for amyotrophic lateral sclerosis (ALS), cancer and age-related diseases, aiming to develop personalized treatments.

Finding depression

Artificial intelligence is also increasingly seen as a means for detecting depression and other mental illnesses, by spotting patterns that may not be obvious, even to professionals.

A research paper by Florida State University's Jessica Ribeiro found it can predict with 80 to 90% accuracy whether someone will attempt suicide as far off as two years into the future.

Facebook uses AI as part of a test project to prevent suicides by analyzing social network posts.

And San Francisco's Woebot Labs this month debuted on Facebook Messenger what it dubs the first chatbot offering "cognitive behavioral therapy" online — partly as a way to reach people wary of the social stigma of seeking mental health care.

New technologies are also offering hope for rare diseases.

Boston-based startup FDNA uses facial recognition technology matched against a database associated with over 8,000 rare diseases and genetic disorders, sharing data and insights with medical centers in 129 countries via its Face2Gene application.

Cautious optimism

Lynda Chin, vice chancellor and chief innovation officer at the University of Texas System, said she sees "a lot of excitement around these tools" but that technology alone is unlikely to translate into wide-scale health benefits.

One problem, Chin said, is that data from sources as disparate as medical records and Fitbits is difficult to access due to privacy and other regulations.

More important, she said, is integrating data in health care delivery where doctors may be unaware of what's available or how to use new tools.

"Just having the analytics and data get you to step one," said Chin. "It's not just about putting an app on the app store." — AFP

Artificial intelligence can improve health care by analyzing data from apps, smartphones and wearable technology — AFP