DeepMind’s Psychlab Heralds Dawn of Artificial General Intelligence

Conversations about AI tend to oscillate between the wildly optimistic and the obnoxiously dystopian. Some pundits will argue we’re on the verge of a new renaissance, in which self-driving cars will spirit us between locations while robotic housekeepers do our bidding. Others foresee a Terminator-like apocalypse. As is often the case, the truth lies somewhere between these extremes. However, no one should be under the illusion that artificial general intelligence (AGI), meaning an AI that can learn the same tasks a human can, is a vague and distant reality. DeepMind has put any doubt to rest with its recent release of Psychlab, a toolkit for assessing artificial intelligence with the same psychological tools we use for assessing human cognitive abilities.

DeepMind is the company behind the algorithm that defeated Lee Sedol in Go. The company has pioneered work on “reinforcement learning” algorithms, which utilize the same general-purpose recipes that underpin much of human and animal learning. In the paper’s introduction, the authors catalog the various accomplishments chalked up by state-of-the-art deep reinforcement learning, which include “navigating 3D virtual worlds viewed from their own egocentric perspective, binding and using information in short term memory, playing ‘laser tag,’ foraging in naturalistic outdoor environments with trees, shrubbery, and undulating hills and valleys and even responding correctly to natural language commands.”

These are all activities humans and our primate cousins engage in, and if this doesn’t read like a catalog of the qualities belonging to an artificial general intelligence, than I don’t know what does. To see a reinforcement learning algorithm in action, check out this YouTube video demonstrating an AI succeeding at the same kind of learning task that is used by psychologists to assess the cognitive skills of rats, primates, and other animals with general intelligence.

Some pundits may refuse to read the writing on the wall regarding AGI because single-purpose, supervised learning algorithms (which possessed no generalizable skills) previously accomplished these tasks. These were the equivalent of one-trick ponies. This is not the case with state-of-the-art deep reinforcement learning – a single algorithm can learn a wide variety of skills just as a single human can. The authors of the Psychlab paper believe these deep reinforcement algorithms can be measured with the same tools that we measure ourselves and other creatures possessing generalized intelligence, such as visual search, change detection, random dot motion discrimination, and multiple object tracking.

While DeepMind has generally been coy regarding the similarity of its work to artificial general intelligence, even playing directly into the hands of naysayers, it will be harder and harder to hide the obvious: A single algorithm can now learn many of the same tasks humans can, and much better in some cases.

Some of the remaining performance gap between ourselves and such algorithms will likely be diminished by improved hardware – sensors and control systems that will allow these algorithms the physical degrees of freedom and sensing power we possess. But these are problems with tractable engineering solutions. Therefore, we shouldn’t surprised when robots begin accomplishing the same tasks previously manned by humans in the workforce.

Whether such AIs will rise up against their human overlords in some epic battle for supremacy remains much in doubt — after all, the reward function used in such reinforcement learning algorithms is not some open-ended mystery, but rather explicitly given by the programmers. But we shouldn’t kid ourselves that the advent of artificial general intelligence with a skill set similar to our own is decades away. It is a reality that is already upon us.



Study: Apple Watch Detects Abnormal Heart Rhythm with 97% Accuracy

According to a study conducted by researchers at the University of California SF in partnership with the team behind the Cardiogram app for Apple Watch, the heart rate monitor on the device can detect abnormal heart rate with an accuracy of 97 percent.

59:45 Beyond the Higgs: What’s Next for the LHC? – with Harry Cliff The Royal Institution 274K views 34:57 Mind Over Matter Is Real: Experiments Reveal! REAL SPIRIT DYNAMICS 56K views 49:23 The Code That Animates Reality – Not For Beginners! REAL SPIRIT DYNAMICS 5.4K views 30:19 We Are Living In A Simulation – New Evidence! REAL SPIRIT DYNAMICS 1.3M views Quantum Computing: The Past, Present & Future. REAL SPIRIT DYNAMICS 68K views Quantum Gravity Documentary Education Channel 4.6K views 2018 Isaac Asimov Memorial Debate: Artificial Intelligence American Museum of Natural History 96K views The Simulation Theory: Quantum Emergence – With Klee Irwin REAL SPIRIT DYNAMICS 5.9K views To Infinity and Beyond: The Accelerating Universe World Science Festival 245K views Emergence Theory: A Layperson’s Guide Quantum Gravity Research 27K views Is Life Quantum Mechanical? – Prof. Jim Al-Khalili The Artificial Intelligence Channel Recommended for you Quantum Gravity Fermilab 250K views A Beginner’s Guide To Quantum Computing Coding Tech 285K views Gravity and Quantum Mechanics – The Quest for Unification | Joe Polchinski aoflex 92K views Our Mathematical Universe with Max Tegmark The Royal Institution 108K views The Simulation Theory: What Is The Reality Code? With Klee Irwin REAL SPIRIT DYNAMICS 81K views Sir Roger Penrose – How can Consciousness Arise Within the Laws of Physics? The Artificial Intelligence Channel Recommended for you The Search for the Theory of Everything – with John Gribbin The Royal Institution 92K views Quantum Craziness | Space News ThunderboltsProject 178K views Imagining The Singularity: What Happens When Computers Transcend Us? The Artificial Intelligence Channel Recommended for you Quantum Gravity: A New Theory Of Everything

58:28 Humanities as a Vocation: Career Paths Beyond the Blackboard University of California Television (UCTV) 204 views New 56:10 Follow Your Gut: Microbiomes and Aging with Rob Knight – Research on Aging University of California Television (UCTV) 65K views 1:23:49 Heart Healthy Eating – Research on Aging University of California Television (UCTV) 7.7K views 1:23:51 Advances in the Treatment and Diagnosis of Acute Stroke: What You Need to Know University of California Television (UCTV) 3.6K views Michio Kaku & Ray Kurzweil – Singularity is Close! Cuckoo for Kaku Recommended for you Examining Heart Failure: How to Recognize and Treat the Weak Heart University of California Television (UCTV) 3.5K views The Aging Face University of California Television (UCTV) 7.9K views An Introduction to Ocular Function and Dysfunction University of California Television (UCTV) 3.7K views New Guidelines for Treatment of Cholesterol: Prevention of Heart Disease and Stroke University of California Television (UCTV) 16K views The Role of Oceans in Our Society and Our Role in Shaping the Future of Ocean Inquiry University of California Television (UCTV) 721 views Before Stroke , Your Body Will Send You These Warning Signs – Don’t Ignore Them ! everyday culture 3.5M views My Aching Neck – My Aching Back University of California Television (UCTV) 2K views MIT AGI: Cognitive Architecture (Nate Derbinsky) Lex Fridman Recommended for you New Steven Pinker – The Philosophy of Free Will The Artificial Intelligence Channel Recommended for you Obesity: Facts and Fictions University of California Television (UCTV) 19K views Principles of a Healthy Diet: How Do We Know What to Eat? University of California Television (UCTV) 121K views SPHERE: Collaborative Research and Precision Medicine – Developmental Disabilities University of California Television (UCTV) 120 views New Dietary Fats: The Good the Bad and the Ugly University of California Television (UCTV) 96K views The Story of Spacetime – with Fay Dowker The Royal Institution Recommended for you Sleep and Aging – Research on Aging University of California Television (UCTV) 18K views Warning Signs of a Stroke

6:02 8 BEST APPS FOR APPLE WATCH SERIES 3 (FEB 2018) Apple Watch World 106K views 16:34 Top 20 Tips and Tricks for the Apple Watch Chaks Corner 69K views 9:54 Best Apps for Apple Watch – Tools Apple Fox 23K views 10:46 Apple March 2018 Event – What to Expect! ZONEofTECH 55K views New Unbelievably Useful Apple Watch Apps DailyTekk 206K views 13 iPhone Settings You Should Change Now! ThioJoe 2.3M views Clean Up Your iPhone or iPad IN UNDER 10 MINUTES!!! David A. Cox 7.6K views New BEST APPLE WATCH APPS 2018 | Must Have Apple Watch Series 3 Apps | Ysis Lorenna Ysis Lorenna 40K views Top 12 Apple Watch ⌚️ Apps- 2018 Jj Rr 3.6K views Best Apps for Apple Watch! Tech IRL 4.1K views 25+ Affinity Photo for iPad beginner’s tips! 9to5Mac 660 views New ODROID-XU4 vs Tinker Board & Raspberry Pi 3 ExplainingComputers Recommended for you Texting on Apple Watch – The Experience Apple Fox 4.7K views Best iPad Apps – Utilities Apple Fox 462 views 10 Things that Annoy me about the Mitsubishi Outlander PHEV Jon Warburton Recommended for you Improve Apple Watch battery life (Tips & Tricks) iAppleGeek 92K views Unbelievably Useful Apple Watch Tips #1 DailyTekk 59K views New Garmin Forerunner 645 vs 935 vs vivoactive 3 Review | Best GPS Running Watches RIZKNOWS 1.4K views New Should You Buy iPhone 6 in 2018? Apple Fox 6K views WWDC 2018 | Rumor Roundup Cult of Mac 477 views New Best Apps for Apple Watch – Communication

Free Trial Ad Up next AUTOPLAY 58:28 Humanities as a Vocation: Career Paths Beyond the Blackboard University of California Television (UCTV) 204 views New 1:30:52 MIT AGI: Cognitive Architecture (Nate Derbinsky) Lex Fridman Recommended for you New 56:10 Follow Your Gut: Microbiomes and Aging with Rob Knight – Research on Aging University of California Television (UCTV) 64K views 58:46 Food and Mood: The Nutrition Connection – Research on Aging University of California Television (UCTV) 24K views How to Talk to Teens: A Developmental Approach University of California Television (UCTV) 4.5K views Engaging Autism: Developmental Implications for Early Intervention University of California Television (UCTV) 1.7K views Sleep Disorders University of California Television (UCTV) 8.5K views Genetics of Autism Spectrum Disorders University of California Television (UCTV) 7.8K views CARTA: The Evolution of Human Nutrition University of California Television (UCTV) 6.7K views Innovations in Dental Care and DentiCal – Development Disabilities University of California Television (UCTV) 94 views New Stem Cells: The Brain’s Beginnings University of California Television (UCTV) 17K views Ian Morris | Why the West Rules — For Now The Oriental Institute 224K views Motor and Social Skill Function: Connecting Learning and Brain Function in Autism University of California Television (UCTV) 3.9K views East West Street: On the Origins of Genocide and Crimes Against Humanity with Philippe Sands University of California Television (UCTV) 294 views New Sexuality and Disability – Developmental Disabilities University of California Television (UCTV) 214 views New Children’s Hospitals and the Importance of Health Insurance and Medicaid University of California Television (UCTV) 112 views New Genes Genomes and Human Disease Part 1 University of California Television (UCTV) 8.9K views Brain Research: New Discoveries and Breakthroughs at UC Davis University of California Television (UCTV) 20K views Childhood Developmental Screening University of California Television (UCTV) 1.3K views Sir Roger Penrose – How can Consciousness Arise Within the Laws of Physics? The Artificial Intelligence Channel Recommended for you SPHERE: Collaborative Research and Precision Medicine – Developmental Disabilities

AI and the Future of Work: Will Our Jobs Disappear?

 Opinions expressed by Forbes Contributors are their own.

Will trucking jobs eventually disappear? Photographer: Bartek Sadowski/Bloomberg

“The sky is falling! The sky is falling!”

Mainstream media has been consistently covering the negative discourse about AI and the future of work, with conversations often led by tech giants like Elon Musk and Bill Gates. We are living in a time that is reflective of the early 2000’s. I remember a world of hysteria and bunkers filled with food, insurance, and so on.

There seems to be something missing within the argument that there will be no jobs in the future due to AI. Elon Musk and Bill Gates are both brilliant and hugely successful CEOs, but neither trait speaks specifically to their foresight. The people we should be turning to to answer this problem are historian and futurists.

One of the most convincing technologists and futurists, Ray Kurzweil, says we have nothing to fear but fear itself. He explains that we have already eliminated all jobs several times over the course of history. This was accomplished with various disruptive technologies, be it the printing press, the steam engine or the automobile. Today’s cycle is neither new nor catastrophic.

In 1900, two-thirds of the population worked in agriculture or manufacturing, with 38% on farms, and 25% in factories; Today, 1 in 10 do, with 2% and 9% respectively. Does that mean that 50% of the population is unemployed? Absolutely not! We have created hundreds and thousands of jobs that never existed.

The same sentiment goes for the future. Yes, the 3.1 million people that drive cars and trucks may eventually have their jobs eliminated. However, these eliminated jobs will certainly be replaced by industries and concepts that are yet to exist.

A job is defined as a piece of work for an agreed upon price. If we look around at the world today, are all of the tasks that need doing close to being accomplished? Not quite. War, poverty, discrimination, inequality, inefficiency, hunger, and slow food delivery… all of these massive problems continue to exist, and these large issues give way to equally large opportunities for people to find work. Does AlphaGo have the solution to these massive and intractable problems? If only that were the case.

An intriguing number of CEOs are convinced that humans will be replaced by machines. Instead of seeing this as inevitable, we need to see this as optional. Executives at blue-chip companies will have a choice to make – focus on cost-cutting or focus on creativity and quality.

Take the example of Accenture. They had 20,000 people employed in a job type that was prone to repeatable activities. Accenture challenged these employees to find a way to use technology to automate parts of their job, and if they did so, they would actually promote them to higher value activities. An amazing 60% of those employees’ jobs were automated away by the employees themselves! Not a single person was made redundant. Instead, they all now work on higher value activities. Contrast this approach with the many companies that believe that if a job is automatable, the person is no longer needed.

Gina Rometti, CEO of IBM, offers a more utopian and nuanced view. She makes the case that jobs will be changed by AI, but not replaced by AI. It is not man or machine. It is, as it has been since the invention of the first stone tool, man and machine. Take physicians, specifically radiologists, as an example. AI now can detect tumors on an X-ray. Can they do it better than humans? In some cases, yes. What the research actually shows that AI is better at detecting some types of tumors, while doctors are better at detecting other types. The solution: Doctor + AI, neither one in isolation. This is at the heart of the term Intelligence Augmentation (IA) rather than Artificial Intelligence (AI).

Whether jobs are destroyed or changed, the more important question is where will these new jobs come from? Again, history teaches us important lessons. In the last few decades, 30 million new jobs were created. They were not created by big enterprise nor by mom-and-pop shops; they were created by high-growth new ventures. Look at the list of highest market cap companies today – only 12% of the companies that existed in 1955 remain today. This is inevitable in the cycle of corporate destruction. For all the talk of technology bubbles, the most successful way to foster job creation is to invest venture money in entrepreneurs.

More recently, the WEF chairman, Klaus Schwab, said that the key traits to the future of work will be things that AI cannot mimic. Creativity. Empathy. Compassion. These are uniquely human traits that no AI guru is claiming are going to be automatable anytime soon. No matter how many movies like HER make us think we will all fall in love with Siri, the fact of the matter is that we can barely make mobile banking and dishwashers automated… it’s a serious disconnect to think that this will happen to the bulk of human existence.

As evident by the number of problems we hear about on the news every day, there is endless work to be done in the modern world. Our task now is not to fret about repetitive activities being automated, but to discover how to apply creative and uniquely human thought to solving these grand problems.