Wednesday, June 29, 2016
AI as "Top Gun" defeats Air Force ace in simulator: "The ALPHA artificial intelligence (AI) created by a University of Cincinnati doctoral graduate is a milestone in the use of genetic-fuzzy systems with specific implementation in unmanned combat aerial vehicles (UCAVs) in simulated air-combat missions. ALPHA's programming involved deconstructing the challenges of aerial fighter deployment into sub-decisions consisting of high-level tactics, firing, evasion, and defensiveness. The language-based fuzzy-logic algorithms cover a multitude of variables, and ease the instilling of expert knowledge to the AI; ALPHA's programming also can be generationally improved. The earliest version of ALPHA consistently beat other AI opponents used by the U.S. Air Force Research Laboratory for research purposes. Subsequent matches against a more mature iteration by a human opponent also proved the AI's invincibility, as retired U.S. Air Force Colonel Gene Lee could not defeat ALPHA, and was consistently bested by the program during protracted clashes in a flight simulator." From http://magazine.uc.edu/
Monday, June 27, 2016
"Prediction is difficult, especially about the future" (attributed with some dispute to Niels Bohr). The World Economic Forum's top ten technologies for the coming year are out (http://www.kurzweilai.net/): nanosensors, next-gen batteries, the blockchain (especially uses beyond BitCoin), 2D materials, autonomous vehicles, organs on chips, new solar cells, open AI, optogenetics, and metabolic engineering.
Monday, June 20, 2016
What is the "deep learning" that was used in the impressive program that defeated a champion Go player? Wired (http://www.wired.com/) article that puts the accomplishment in perspective and provides a high-level description of deep learning (in 6 brief steps). Placing it in perspective: "Thus, deep learning (and machine learning in general) has proven to be a powerful class of methods in AI, but current machine learning methods require substantial human involvement to formulate a machine learning problem and substantial skill and time to iteratively reformulate the problem until it is solvable by a machine." (Unfortunately the author dismisses the field of Genetic Algorithms as a failure, and I take exception to that -- GA can do amazing things, but just because it has limitations doesn't make it a failure.)
Wednesday, June 15, 2016
New research from Yale uses sound to amplify light on a silicon chip. That likely sounds esoteric and meaningless, but the Internet runs on a backbone of optical cables. When signals get routed they currently convert signals to electricity to determine the routing and convert signals back to light to run on an optical cable. Logic gates using light have been invented, but with no way to amplify the light they are of little use. This invention could remove that inefficient conversion to electricity leading to improvements on the backbone of the Internet. Details are here: http://news.yale.edu/ (There is a way to amplify light in an optical cable using lasers but I don't think that technique works on a chip. Passive optical amplifiers ref: http://www.fiberopticshare.com/)
Friday, June 10, 2016
Can a computer observe online behavior and spot when someone becomes radicalized? Researchers at Lancaster U. in the UK think they can. They looked at 100 million tweets across 100K accounts. The research shows that when users begin either sharing tweets from known pro-ISIS accounts, or using extremist language they quickly display a large change in the language they use, tweeting new words and terms, and indicating a clear shift in online behavior. http://www.lancaster.ac.uk/
Wednesday, June 8, 2016
Seemingly unrelated information can be usefully mined. In this case, Microsoft has found that examining search queries allows them to identify pancreatic cancer patients before they even know it themselves! http://www.nytimes.com/ What may not be immediately obvious is that a malicious entity examining search queries may be able to glean information for nefarious purposes.
How does a computer play chess? This website, http://www.bewitched.com/chess/, shows the moves that the computer is exploring before deciding on its move. Basically it is trying many moves ahead, putting a score on each move, and choosing the best move. In general, given two computers playing against each other, the one which can look one move further ahead will win more than it loses (assuming that the scoring system is relatively equal). Knowledge of chess shows in the scoring algorithm. Give it a try!
Tuesday, June 7, 2016
Robot-assisted surgery exists and will likely increase. This http://spectrum.ieee.org/ video looks like a compilation of ads, but it is still interesting. Note that medical software isn't without risks -- see the classic Therac-25 case (https://en.wikipedia.org).
Sunday, June 5, 2016
A VISA ring instead of a card will be tested at the Rio Olympics: http://www.engadget.com/. Unfortunately, technical details are missing from this article except for this comment which does not make one comfortable: "It doesn't exchange as much data as Apple Pay or Android Pay, but it's on par with swiping your card." Convenience at the expense of security.
Over the years there have been a variety of attacks speculated for malicious hardware including one that manipulates the random number generator to break encryption. Here is a particularly insidious one because of the tiny change (add a capacitor) and a demonstration of the backdoor that is created. The paper appears in the IEEE Symposium on Privacy and Security, but this https://www.wired.com/ article provides a brief description.
Thursday, June 2, 2016
Mario Brothers is hard -- in a computational sense. That is, figuring out the optimal strategy is hard (according to research at MIT). Better understanding where a game fits in computational theory puts us closer to understanding how computationally hard real life is. These are questions probed in a computational theory course (such as CSE 460 at MSU).
Wednesday, June 1, 2016
Apologies for a post related to politics, but this sysadmin view of the Clinton email mess explains it better than any political posting I've seen. It describes the "entitled executive," a creature that exists in the corporate word. Hillary fits: http://arstechnica.com/