IBM's Doctor Watson?

From USATODAY.com

"IBM's supercomputer system, best known for trouncing the world's best "Jeopardy!" players on TV, is being tapped by one of the largest U.S. health insurers to help diagnose medical problems and authorize treatments.

WellPoint Inc., which has 34.2 million members, will integrate Watson's lightning speed and deep health care database into its existing patient information, helping it choose among treatment options and medicines."

The WellPoint application will combine data from three sources: a patient's chart and electronic records that a doctor or hospital has, the insurance company's history of medicines and treatments, and Watson's huge library of textbooks and medical journals.

IBM says the computer can then sift through it all and answer a question in moments, providing several possible diagnoses or treatments, ranked in order of the computer's confidence, along with the basis for its answer.

Previously IBM developed Deep Blue, "a massively parallel computer, to be applied to the study of biomolecular phenomena such as protein folding."

Stanford offers free CS courses

From DeviceGuru

Last year at Educause, I heard Neil Gershenfeld (of MIT's Center for Bits and Atoms) speak on a shift away from a scarcity-based model of education. A significant point was that technology has the potential to radically revolutionize the business of education. In the past, access to ideas was limited by access to physical resources: books, laboratories, faculty members. When resources are digitized, the nominal cost of reproducing them becomes insignificant.

In the latest of a series of similar examples, Stanford University has announced that it will be offering a series of 10 free online computer science courses, including their introduction to computer science, artificial intelligence, and robotics. As of a couple of days ago, more than 120,000 students had signed up for the courses.

'Software is the Modern Language of Science'

From HPCwire

Computation, according to Nobel physics laureate Ken Wilson, has joined theory and experimentation as a "third leg" of scientific discovery (see Denning 2007, for example).

“We now have very small periods in time that are leading to very large changes in the amount of data, the amount of computation, and the amount of knowledge that is needed in order to carry out this kind of work,”

“...after more than four centuries of science being conducted at a painstakingly slow pace, today’s communications technologies and scientific advances are forcing a dramatic change–and acceleration–in all areas of science. At the heart of this change will be software.”

More recently, some have even been talking of a "fourth paradigm" -- the use of massive databases to discover new knowledge. For more see "The Fourth Paradigm," a series of essays published by Microsoft Research.

Robots learning from experience

See original at ICT Results

Great example of "bottom-up" approaches to AI, which rely on statistical learning techniques to acquire knowledge, rather than attempting to hand-code it all explicitly.

"Software that enables robots to move objects about a room, building up ever-more knowledge about their environment, is an important step forward in artificial intelligence.

Some objects can be moved, while others cannot. Balls can be placed on top of boxes, but boxes cannot be stably stacked on top of balls. A typical one-year-old child can discover this kind of information about its environment very quickly. But it is a massive challenge for a robot – a machine – to learn concepts such as ‘movability’ and ‘stability’, according to Björn Kahl, a researcher at the Bonn-Rhein-Sieg University and a member of the Xpero robotics research project team.

The aim of the Xpero project was to develop a cognitive system for a robot that would enable it to explore the world around it and learn through physical experimentation."

Grand unified theory of AI: Combining top-down with bottom-up

See original at Science Daily

In the 1950s and '60s, artificial-intelligence researchers saw themselves as trying to uncover the rules of thought. But those rules turned out to be way more complicated than anyone had imagined. Since then, artificial-intelligence (AI) research has come to rely, instead, on probabilities -- statistical patterns that computers can learn from large sets of training data.

The most well-known example of the top-down approach to AI is the Cyc project. I wonder how it could be combined with bottom-up approaches?

At The Extreme Edge Of Artificial Intelligence - Forbes.com

See original at Forbes.com

Think sentient computers are the stuff of science fiction? We have news for you.

The last couple of posts were a bit pessimistic; here's someone writing from the other side. Make sure to check out the link to 11 Leaders in Artificial Intelligence.

Computers: Will They Ever Learn?

See original at Forbes.com

Have we plateaued in our progress on AI?

A computer that translates as well as a human, which is as good a definition as any of genuine artificial intelligence, is nowhere in sight.

This isn't just an issue for computer translation; it exists for just about all of the tough problems computers are working in: language, vision, robotics--even search, which is at the heart of the Web. Computers have become remarkably useful but are still a far cry from the dream of early computer pioneers of a machine so smart it would be indistinguishable from a human.

AI Report

The AI Report - Forbes.com

"Can machines think? In 1950, Alan Turing, considered by some to be the father of modern computing, published a paper in which he proposed that, "If, during text-based conversation, a machine is indistinguishable from a human, then it could be said to be 'thinking' and, therefore, could be attributed with intelligence." He predicted that a computer would pass this "Turing Test" by the end of the century. That hasn't happened--yet. But the question continues to provoke and inspire. AI might be just around the corner, or it might be centuries away"

GR: check out this Forbes special report, with essays by cyberneticist Kevin Warwick, philosopher Nick Bostrom, Singularity Institute president Michael Vassar, and Google Research Director Peter Norvig.

By 2040 you will be able to upload your brain...

Upload your brain... "By 2040 you will be able to upload your brain... or at least that's what Ray Kurzweil thinks. He has spent his life inventing machines that help people, from the blind to dyslexics. Now, he believes we're on the brink of a new age – the 'singularity' – when mind-boggling technology will allow us to email each other toast, run as fast as Usain Bolt (for 15 minutes) – and even live forever. Is there sense to his science – or is the man who reasons that one day he'll bring his dad back from the grave just a mad professor peddling a nightmare vision of the future?"

"In Kurzweil's estimation, we will be able to upload the human brain to a computer, capturing "a person's entire personality, memory, skills and history", by the end of the 2030s; humans and non-biological machines will then merge so effectively that the differences between them will no longer matter; and, after that, human intelligence, transformed for the better, will start to expand outward into the universe, around about 2045. With this last prediction, Kurzweil is referring not to any recognisable type of space travel, but to a kind of space infusion. "Intelligence," he writes, "will begin to saturate the matter and energy in its midst [and] spread out from its origin on Earth."

An intelligent system avoids forgetting things

Artificial Intelligence applied to help care for the aging: "An elderly lady is about to go to bed. She goes into her room, sits down on the bed, takes off her slippers and turns off the light. Suddenly, before getting into bed, a small alarm goes off and a mobile device reminds her that she has not taken her tablets.

This is how the new intelligent system developed by researchers from the Department of Computer Science and Artificial Intelligence of the UGR works. Maria Ros Izquierdo is from the Higher Technical School of Computer Engineering of the UGR and the co-author of a study which is published this month in the Expert Systems with Applications magazine. 'It is a prototype which, in a non-intrusive manner, facilitates the control of the activity of people with special needs and increases their independence', she explained to SINC."

"A team of researchers from the University of Granada (UGR) has created a system with Artificial Intelligence techniques which notifies elderly people or people with special needs of the forgetting of certain everyday tasks. This system uses sensors distributed in the environment in order to detect their actions and mobile devices which remind them, for example, to take their keys before they leave home."

Machines computing morality

Moral machines: "Researchers from Portugal and Indonesia describe an approach to decision making based on computational logic in the current issue of the International Journal of Reasoning-based Intelligent Systems which might one day give machines a sense of morality."

"Morality no longer belongs only to the realm of philosophers. Recently, there has been a growing interest in understanding morality from the scientific point of view," the researchers say.

"They have turned to a system known as prospective logic to help them begin the process of programming morality into a computer. Put simply, prospective logic can model a moral dilemma and then determine the logical outcomes of the possible decisions. The approach could herald the emergence of machine ethics.

"The development of machine ethics will allow us to develop fully autonomous machines that can be programmed to make judgements based on a human moral foundation. "Equipping agents with the capability to compute moral decisions is an indispensable requirement," the researchers say, "This is particularly true when the agents are operating in domains where moral dilemmas occur, e.g., in healthcare or medical fields."

Researchers develop "robotic apprentices"

PC Pro: Researchers develop "robotic apprentices": "University researchers have developed an artificial intelligence that can learn by watching 'experts' perform a task.

The artificial intelligence was developed by Stanford University for use in robotic helicopters which learn to fly and perform stunts by watching 'expert' helicopters perform the same tasks, rather than by having software engineers input every individual instruction."

'Frankenrobot' Has Biological Brain

Discovery News "Meet Gordon, probably the world's first robot to be controlled exclusively by living brain tissue.

"Stitched together from cultured rat neurons, Gordon's primitive grey matter was designed at the University of Reading by scientists who unveiled the neuron-powered machine on Wednesday.

"Their groundbreaking experiments explore the vanishing boundary between natural and artificial intelligence, and could shed light on the fundamental building blocks of memory and learning, one of the lead researchers said."

Are you a human scanner?

Scientific American "Those anti-bot security forms that slow you down when you're entering information just might serve a larger purpose. You're just about ready to buy a pair of tickets on Ticketmaster, but before you can take the next step, an annoying box with wavy letters and numbers shows up on your screen. You dutifully enter in what you see—and what a bot presumably can't—in the name of security."

"But what you may not know is that you also have helped archivists decipher distorted characters in old books and newspapers so that they can be posted on the Web."

A sign of human-machine collaborative thinking?

Robot swarms 'evolve' effective communication

New Scientist Tech: "Robots that artificially evolve ways to communicate with one another have been demonstrated by Swiss researchers. The experiments suggest that simulated evolution could be a useful tool for those designing of swarms of robots.

Roboticists Dario Floreano, Sara Mitri, and Stéphane Magnenat at the Swiss Federal Institute of Technology in Lausanne collaborated with biologist Laurent Keller from the University of Lausanne.

They first evolved colonies of robots in software then tested different strategies on real bots. Both simulated and real robots were set loose in an arena containing two types of objects – one classified as 'food' and another designated 'poison' – both lit up red.

In each experiment, five hundred generations were evolved this way, under different selective pressures. "Under some conditions, sophisticated communication evolved," says Keller. "We saw colonies that used their lights to signal when they found food and others that used signals to communicate they had found poison."

Computer Model Mimicks How Brain Recognizes Street Scenes

ScienceDaily — "At last, neuroscience is having an impact on computer science and artificial intelligence (AI). For the first time, scientists in Tomaso Poggio's laboratory at the McGovern Institute for Brain Research at MIT applied a computational model of how the brain processes visual information to a complex, real world task: recognizing the objects in a busy street scene. The researchers were pleasantly surprised at the power of this new approach."