| - MACHINE LEARNING - General Index by Topic to AI in the news |
AI Topics Home | ||||
|
October 29, 2007: The Semantic Web Goes Mainstream - Radar Networks is unveiling a new tool that provides a smarter way to find information and increase productivity. By Kate Greene. Technology Review. "Radar Networks, based in San Francisco, is releasing a free Web-based tool, called Twine, that it hopes will change the way people organize their information. Twine is a website where people can dump information that's important to them, from strings of e-mails to YouTube videos. Or, if a user prefers, Twine can automatically collect all the Web pages she visited, e-mails she sent and received, and so on. Once Twine has some information, it starts to analyze it and automatically sort it into categories that include the people involved, concepts discussed, and places, organizations, and companies. This way, when a user is searching for something, she can have quick access to related information about it. ... The idea underlying Twine's function and technologies is known as the Semantic Web, a concept, long discussed in research circles, that can be described as a sort of smart network of information in which data is tagged, sorted, and searchable. ... In addition to employing the Semantic Web standards, Twine is also using extremely advanced machine learning and natural-language processing algorithms that give it capabilities beyond anything that relies on manual tagging. ... Twine will open up to invited users starting today." October 26, 2007: Video search makes phone a 'second pair of eyes'. By Will Knight. NewScientist.com news. "Soon, however, it may be easier to simply record a video clip of an item of interest and have your phone tell you about it instead. Researchers at Accenture Technology Labs in France have developed technology that makes this possible using any ordinary 3G cellphone equipped with a video camera. ... If a user records a video clip of, say, a foreign food item, the system can automatically identify ingredients that might cause an allergic reaction. Similarly, when shown a book, it can quickly perform an online price comparison, or find a review (see video...). Live video footage is fed from the handset to a central server, which rapidly matches on-screen objects to images previously entered into a database. The server then sends find relevant information and sends it back to user. The central server uses an algorithm called the Scale-Invariant Feature Transform (SIFT) to match objects. ... Microsoft has a system called Lincoln, that lets users to take snapshots and send them off for identification. Another system developed by Evolution Robotics of Pasadena, California, called ViPR, also uses video footage to identify objects, and is already available in Japan."
>>> Image Understanding, Information Retrieval, Machine Learning, Vision, Applications October 25, 2007: Rating Facial Expressions - New software could help mental-health professionals assess patients and ensure that salespeople project a positive attitude. By Anna Davison. Technology Review. "Software that recognizes and rates smiles was demonstrated recently at an exhibition in Tokyo, where attendees competed to outsmile one another. The smile-checking technology is the latest addition to Omron Corporation's OKAO Vision software suite, which detects faces in images and can determine the person's gender and approximate age, or verify his or her identity from a database of faces. The smile software is Omron's first foray into facial-expression detection and analysis, a field that could revolutionize how humans interact with machines, and with each other. ... 'Clearly, it's an interesting thing,' says Joseph Atick of L-1 Identity Solutions, based in Stamford, CT, which supplies identification technology, primarily for security applications. 'If you can read people better, you can serve them better.' ... Sophisticated facial-expression analysis could help mental-health professionals evaluate their patients and monitor their progress." October 22, 2007: 'Smart' video offers an alert to threats - Taking boredom factor out of security systems. By Hiawatha Bray. The Boston Globe. "In video surveillance systems, the weakest link is the often bored, distracted human who has to spend hours staring at a bank of video monitors, waiting for something suspicious to happen. Several Boston area companies say they have found a solution: surveillance systems smart enough to recognize threats, even when their human operators do not. 'It essentially replaces the need for people to watch video,' said Scott Schnell, chief executive of VideoIQ Inc., a Bedford firm that was spun off earlier this year from General Electric Co. ... Systems from VideoIQ and Intuvision Inc. of Woburn can automatically spot an intruder climbing a fence or a subway passenger leaving a suspicious parcel on the platform. ... [Simon] Harris said that worldwide sales of smart video surveillance systems will be less than $100 million this year, but rise to about $3 billion by 2010. ... One test video shows ducks and boats on the Hudson River. The system draws yellow boxes around the harmless ducks, but when a boat appears, the box turns bright red. ... Intuvision, a startup funded by grants from the US intelligence community, has attacked the problem using a technique called 'task-based attention.'"
>>> Law Enforcement, Image Understanding, Vision, Machine Learning, Applications, Industry Statistics October 19, 2007: What I Meant to Say Was Semantic Web. John Markoff's post to Bits, The New York Times' Technology Blog. "One great way to start a fight in a crowded Silicon Valley cocktail party (and there are a lot of them these days) is to mention Web 3.0. There is no easy consensus about how to define what is meant by Web 3.0, but it is generally seen as a reference to the semantic Web. While it is not that much more precise a phrase, the semantic Web refers to technology to make using the Internet better by understanding the meaning of what people are doing, not just the way pages link to each other. ... So companies are bubbling up all over the place that claim to be building part of the semantic Web. Some are building voice recognition systems to use while browsing the Internet on a cell phone. Some want to challenge Google head on with a better search engine. ... In a demonstration I saw earlier this week Twine appeared to do a good job of what artificial intelligence researchers refer to as 'entity extraction,' that is categorizing things like people and places automatically."
>>> Interfaces, Representation, Web-Searching Agents, Natural Language Processing, Machine Learning, Knowledge Management, Applications October 11, 2007: 'Dark Web' Project Takes On Cyber-Terrorism. By Steven Kotler. FOXNews.com. "'Since the events of 9/11, terrorist presence online has multiplied tenfold,' says Hsinchun Chen, director of the University of Arizona's Artificial Intelligence Lab. 'Around the year 2000, there were 70 to 80 core terrorist sites online; now there are at least 7000 to 8000.' Those sites are doing everything from spreading militant propaganda to offering insurgency advice to plotting the next wave of attacks, making the net, as Chen also points out: 'arguably the most powerful tool for spreading extremist violence around the world.' But thanks to Chen, that tide may be turning. He's the architect behind the newest weapon in the war on terror -- a giant, searchable database on extremists known as Dark Web. Using a bevy of advanced technologies, Dark Web is an attempt to uncover, cross-reference, catalogue and analyze all online terrorist-generated content. ... Dark Web is Chen's second foray into online crime-fighting. The first began in 1997, when he -- already an expert at tracking social change online (crime and terrorisms being extreme examples of social change) -- teamed up with the Tucson Police Department and the National Science Foundation (NSF) to help develop Coplink, a way for law enforcement forces around the country to link files and consolidate data. ... [Dark Web] utilizes existing technologies... as well as brand new technologies like sentiment analysis, which is capable of scanning documents for emotionally charged keywords such as 'that sucks.'... Civil-liberties concerns may continue to dog the technological front of the war on terror, but Dark Web is already producing results." October 5, 2007:The 'Numb3rs' Don't Lie [radio broadcast]. NPR's Talk of the Nation: Science Friday, with Ira Flatow. "Mathematics may seem like an unusual tool to catch criminals, but real math and actual events inspire the CBS crime drama Numb3rs. Guests [Gary Lorden & Keith Devlin] discuss the intersection of math-based crime solving and prime-time television. ... [17:15] Flatow: Mm-hmm. Let’s talk a bit about something that you write about in the book. You write that - it has to do with the war on terror. And we know that the government has all kinds of data mining that it’s doing. And you write that machine learning is, quote, 'perhaps the single most important tool within the law enforcement community’s data mining arsenal when it comes to profiling, enhanced catching or preventing criminals and terrorists.' Can you tell us what machine learning is? Dr. Devlin: Okay. That’s - actually, the center that I direct at Stanford is actually the world’s leader in doing that thing. It’s where you - it’s a branch of what was known as artificial - still is known as artificial intelligence. It means you have a computer program which you present lots of data, it could be data about - an obvious one is can you determine the profile of someone entering the country who’s likely to be a terrorist? ..." October 5, 2007: Technology Would Help Detect Terrorists Before They Strike. Press release from the University at Buffalo. " Computer and behavioral scientists at the University at Buffalo are developing automated systems that track faces, voices, bodies and other biometrics against scientifically tested behavioral indicators to provide a numerical score of the likelihood that an individual may be about to commit a terrorist act. 'The goal is to identify the perpetrator in a security setting before he or she has the chance to carry out the attack,' said Venu Govindaraju, Ph.D., professor of computer science and engineering in the UB School of Engineering and Applied Sciences. ... 'We are developing a prototype that examines a video in a number of different security settings, automatically producing a single, integrated score of malfeasance likelihood,' he said. A key advantage of the UB system is that it will incorporate machine learning capabilities, which will allow it to 'learn' from its subjects during the course of a 20-minute interview. That's critical, Govindaraju said, because behavioral science research has repeatedly demonstrated that many behavioral clues to deceit are person-specific." October 5, 2007: From stick figures to artificial intelligence. PhysOrg.com [ source: University of Alberta]. "Today, [Brian] Tanner, 27, is a computer scientist working toward his PhD at the University of Alberta in the area of artificial intelligence and, more specifically, in the area of reinforcement learning. ... He is working at a highly abstract, theoretical level to come up with a mathematical model that will enable a computer to learn how to make its own decisions. His goal is to make such a model a template for all AI applications. ... In addition to being recognized as a 2007 Honorary Izaak Walton Killam Memorial Scholar for his work, Tanner has also been awarded the Dorothy J. Killam Memorial Graduate Prize as the most outstanding Killam recipient in the areas of Engineering, Mathematics and Physical Sciences. ... Prizes such as the U of A Killam awards program both attract and reward such scholars who are doing important work for the betterment of all." October 3, 2007: Don’t invent, evolve - The inventor’s trial-and-error approach can be automated by software that mimics natural selection. Economist.com. "Evolutionary design, as it is known, allows a computer to run through tens of millions of variations on an invention until it hits on the best solution to a problem. As its name suggests, evolutionary design borrows its ideas from biology. It takes a basic blueprint and mutates it in a bid to improve it without human input. As in biology, most mutations are worse than the original. But a few are better, and these are used to create the next generation. ... What has changed, in this as in so much else, is the availability and cheapness of computing power. According to John Koza of Stanford University, who is one of the pioneers of the field, evolutionary designs that would have taken many months to run on PCs are now feasible in days. The result is that the range of applications to which the principles of evolutionary design are being applied is growing fast. ... Perhaps the most cunning use of an evolutionary algorithm, though, is by Dr Koza himself. His team at Stanford developed a Wi-Fi antenna for a client...." October 3, 2007: 3-D avatar to help doctors visualize patient records and improve care. KurzweilAI.net. "IBM's Zurich Research Lab has developed an avatar to allow doctors to visualize patient medical records."
>>> Medicine, Machine Learning, Applications October 3, 2007: Scans reveal lost gravestone text. By Cristina Jimenez. BBC News. "Scientists at Carnegie Mellon university are making high resolution 3D scans of tombstones to reveal the carved patterns in the stone. A computer matches the patterns to a database of signature carvings which reveals the words. The technique could one day also be used by doctors to examine a patient's tongue for signs of illness. ... 'This technology is expected to reduce guessing work in field inspection,' said Dr Yang Cai, director of the Ambient Intelligence Lab at Carnegie Mellon Cylab. ... The new technique allows them to define patterns of 'typical' lines and curves and store them in a database. 'If the computer finds the data matches the patterns in the database, then it will highlight the area,' Dr Cai said." October 2007: Meet the Innovators - The Player: Luis von Ahn's secret for making computers smarter? Get thousands of people to take part in his cunning online games. By Polly Shulman. Smithsonian Magazine [part of the special report: Meet the Innovators . "What excites researchers about [Luis] von Ahn's 'human computation' work, as he calls it, is less the prospect of getting people to accomplish boring, repetitive chores than the promise of training computers to do the chores themselves. Many tasks that are easy for people are surprisingly difficult for computers, especially those that children learn easily, such as classifying objects, recognizing faces, learning verbal languages and reading handwriting. 'We're biologically programmed to teach our kids,' says Manuel Blum, a Carnegie Mellon computer scientist and von Ahn's former adviser. 'We don't have the patience to teach computers the same way, by answering question after question.' Michael Kearns, a computer scientist at the University of Pennsylvania, says, 'There are lots of people studying the hard problem of teaching computers to learn, and lots of other people seeing the entertainment value of the Web. But it's rare to find somebody like von Ahn, who has thought deeply about how to combine the two.' ... His 'big goal,' von Ahn says, is to make computers able to do anything that people can do. 'I think it'll happen, definitely. If not in 50 years, then 100.'"
>>> Machine Learning, Applications October 2007 [issue date]: Cracking GO - Brute-force computation has eclipsed humans in chess, and it could soon do the same in this ancient Asian game. By Feng - Hsiung Hsu. IEEE Spectrum Online. "In 1957, Herbert A. Simon, a pioneer in artificial intelligence and later a Nobel Laureate in economics, predicted that in 10 years a computer would surpass humans in what was then regarded as the premier battleground of wits: the game of chess. Though the project took four times as long as he expected, in 1997 my colleagues and I at IBM fielded a computer called Deep Blue that defeated Garry Kasparov, the highest-rated chess player ever. You might have thought that we had finally put the question to rest -- but no. Many people argued that we had tailored our methods to solve just this one, narrowly defined problem, and that it could never handle the manifold tasks that serve as better touchstones for human intelligence. These critics pointed to weiqi, an ancient Chinese board game, better known in the West by the Japanese name of Go, whose combinatorial complexity was many orders of magnitude greater than that of chess. Noting that the best Go programs could not even handle the typical novice, they predicted that none would ever trouble the very best players. Ten years later, the best Go programs still can't beat good human players. Nevertheless, I believe that a world-champion-level Go machine can be built within 10 years, based on the same method of intensive analysis -- brute force, basically -- that Deep Blue employed for chess. I've got more than a small personal stake in this quest. At my lab at Microsoft Research Asia, in Beijing, I am organizing a graduate student project to design the hardware and software elements that will test the ideas outlined here. ..." FOR MORE ARTICLES, SEE THE MACHINE LEARNING NEWS ARCHIVE |
|||||