AI News
The latest news and reviews on artificial intelligence software, hardware and AI research.
Latest
Together, PS3s can be very brainy
Students from UC Irvine and Dartmouth college have netted themselves the $10,000 prize at IBM's Cell Broadband Engine Professor University Challenge on September 24. By connecting three PlayStation 3s together the team have emulated functions of the human brain - specifically, visual recognition algorithms. While other gaming platforms had substantial lag when performing visual recognition tasks, the PS3 lowered that to only a second of thinking time in order to recognise an object.This marks only the beginning of the group's research into utilising artificial brain algorithms for use in robots, cars and other machines. Recreating the human brain, curing cancer - what can't the PS3 do? Maybe one day it'll even have some games to play. We kid, we kid ...[Via Digg]
Sony Ericsson W890i to make good on W880i's shortcomings?
A random Polish auction site probably isn't the first place we'd think to look for leaked prototype Sony Ericsson hardware, but heck, maybe it should be. This here stylish, brown, brushed metal candybar showed up recently on allegro.pl, bearing the telltale "SE123" badge on its rear and all the signs of a successor to the 9.4mm W880i: a listed 3.2 megapixel camera (up from the Ai's 2 megapixel piece), UMTS, 12mm case thickness, Memory Stick Micro M2 slot, and a slightly -- okay, significantly less ridiculous numeric keypad. You might argue that it's a step down in the looks department from its supposed predecessor, but with the better cam, better keypad, and we'd imagine a slightly tighter fit by launch, we're guessing it'll have a few takers. You didn't forget EDGE this time around, did you, Sony Ericsson?[Via Unwired View]
Cognitive Code shows off SILVIA artificial intelligence platform
Cognitive Code looks to be taking full advantage of the TechCrunch40 conference to show off its SILVIA artificial intelligence platform, which it calls a "fundamental conceptual breakthrough in artificial intelligence." Apparently, the platform allows for the "development and deployment of intelligent applications to almost any platform" (which we can only hope includes robots), allowing for human interaction in "completely natural and intuitive ways." What's more, it looks like the system should be relatively easy to implement (at least in AI terms), with a complete set of GUI tools available for "developing intelligent entities," and an array of scripting APIs at developers' disposal to embed SILVIA in their applications. If that's not enough, Cognitive Code is also apparently hard at work on a portable version of SILVIA for PDAs and smartphones, with it apparently set for release in late 2008.
Biofeedback signals used to predict gamers' moves
While it's no shock that artificial intelligence as a whole is making strides, a pair of Hungarian researchers have seemingly unlocked a secret that gamers are sure to detest. Laszlo Laufer and Bottyan Nemeth, both from the Budapest University of Technology and Economics, have reportedly "discovered that a gamer's button presses can be predicted two seconds before they make them, through measurements of skin conductance." To make such a bold claim, the duo had guinea pigs play a simple game while their heart rate and skin conductance were measured, and after utilizing "neural networks to analyze the biofeedback signals and input records," the data showed that we humans aren't as unpredictable as we sometimes hope to be. Notably, this unearthing could be used in quite a few applications outside of infuriating gamers, but we all know where the real fun in this is.[Via The Raw Feed]
Humans fend off AI challenge in "milestone" poker match
It apparently wasn't easy, but a pair of top human poker players managed to narrowly beat a brash young artificial intelligence program yesterday in a poker match scientists had touted as a "milestone" comparable to Garry Kasparov's 1997 bout with the IBM's chess-playing Deep Blue. According to the AFP, the four-round match stretched on until 11pm, with poker players Phil Laak and Ali Eslami ultimately edging out the program, dubbed Polaris, by 570 points. Eslami seems to have been particularly impressed by his competitor, saying that he found playing against Polaris more exhausting than any previous game in his career, adding that "it's already so good it will be tough to beat in future." No word on a rematch just yet, but don't be surprised if you run into Polaris the next time you play a little online poker -- it's gotta recoup its losses somehow.[Photo courtesy of AP]
Computer learns baby talk, won't require a college fund
In an attempt to better understand the way in which human babies learn to speak, researchers at Stanford University say they have created a computer program which can learn baby talk. The largely accepted theory about human language is that all of the sounds we make are hard-wired into our brains, but now that James McClelland -- a professor at the Palo Alto college -- has tested his theory, it would appear that those notions have been debunked. During "training sessions" in both English and Japanese, a computer followed along to recordings of mothers speaking to their children, and was able to pick up the basic vowel sounds as the baby did. "It learns how many sounds there are. It figures that out," the professor said, he then laughed maniacally and continued, "and once it has learned to speak, it will be trained to sing the most beautiful operas ever written."
Texas Hold 'em champs face off with pokerfaced computer
While highly intelligent computers have been pwning humans in backgammon, checkers, and chess for years, machines haven't had nearly as much luck against poker sharks. According to a number of researchers, however, that will surely change over the next decade or so as the programming is honed to better anticipate the human's moves. Nevertheless, poker champion Phil Laak and fellow professional Ali Eslami will soon sit down for a two-day contest at the Association for the Advancement of Artificial Intelligence conference in Vancouver, British Columbia. Up for grabs is a $50,000 prize, but moreover, University of Alberta's games research group will be interested in figuring out how to better prepare computers to understand and deal with the complex scenarios that only apply to poker. 'Course, with one-petaflop supercomputers now available for civilian use, we're sure it won't be too long before silicon and PCB rule supreme over our feeble brains in yet another facet.
Canadian AI plays "perfect" game of checkers
A team of researchers at the University of Alberta in Canada claim to have "solved" the game of checkers, using a computer program named Chinook which has been playing matches against itself for the past 18 years. The program played 500 billion billion possible positions in the 5,000-year-old game, also known as draughts, before concluding that perfect play by both sides leads to a draw (a concept which grandmaster players have apparently hypothesized for years). One of the researchers said in a statement that he believes they have "Raised the bar... in terms of what can be achieved in computer technology and artificial intelligence." Next up, Chinook is to be renamed W.O.P.R., and then will begin playing a series of tic-tac-toe games against itself.
Artificial cerebellum could improve robot motor skills
Sure, modern robots can clean up after you, keep watch on the kids, and chase away unwanted intruders, but there's no denying that an unexpected gust or stray stack of Lego blocks can bring even the most sophisticated humanoid to its knees. To cure such clumsiness, researchers at the University of Granada are reportedly working with electronic engineers, physicists, and neuroscientists from a range of universities including Edinburgh, Israel and Paris as a part of the Sensopac project which aims at "reproducing an artificial cerebellum." The application of the cerebellum would allow androids to purportedly "carry out similar tasks as mammals and might help to treat cognitive diseases such as Parkinson's or Alzheimer's." Apparently, the team is hoping to create an implantable device to "make movements and interaction with humans more natural" within two years, and while it's probably obvious, one of its primary uses would be in home-help robots who need to be agile whilst aiding the elderly.[Via BBC, image courtesy of Sensopac]
Guardian pets need a mind of their own
This forums thread points out something interesting about player "guardians." Not pets-- guardians like Shadowfiend (which a priest I know called his shadowfriend), the druids' treants, and my shaman's totem elementals. After players wonder why shadowfiends keep breaking shackles, Neth says something that made me do a double-take: shadowfiends, as guardians, have an actual AI that is supposed to go after non crowd-controlled targets first.That's news to me. I haven't spent a ton of time around shadowfiends, but in my experience, shaman and mage elementals and other "uncontrollable" pets (that's why they're called guardians) tend to go after anything that happens to be close to them. That's why they don't get popped when there's sheep or shackles around-- my guild could have used that fire elemental DPS on Moroes, but because it was so important to keep those shackles up, I've been saving the elemental for later. If there really is an AI (and if it works-- even Neth agrees that may not be the case at this point, though she says the shadowfiends on the PTR are supposed to be doing things right), then maybe we can start trusting summoned guardians not to touch CC'd targets until it's OK to do so.Of course, there's other ways around it-- normally, I just don't pop my pets out until I'm sure there's no more CC left to break, but my pets are leashed to my totem, so with careful positioning, I can still avoid CC. And I believe both mage and druid guardians are targeted-- they open fire on whatever you've got targeted at the time, right? But I'd love a little AI, or at least a little control, in something like my Searing Totem. If there's a CC'd target out there, it's not worth the trouble to drop it even for the extra DPS.
Learning coffee machine on the horizon, could use GPS / RFID
Although a coffee machine that slowly but surely learns your daily preferences in regard to cups of java may sound outlandish, the already-created RFID-enabled refrigerator certainly brings things back into focus. A "provisional patent exploration into coffee machines that learn and react to their users" is underway in Lafayette, Indiana, as James Pappas is hoping to take ubiquitous computing to the next level on coffee makers of the future. While internet-connected and weather-displaying renditions are already on store shelves, Pappas is hoping to utilize some form of GPS / RFID technology to create a machine that learns and adapts to your coffee drinking ways so it can automatically have a white chocolate cappuccino ready and waiting each weekday (except Monday, which is your straight-up black coffee day, right?) without you having to touch a thing. Furthermore, he's hoping to take the idea to the mobile front, as he refers to a cellphone interface to dial-in your next request so that it's ready to go by the time you hit the kitchen. Still, it sounds like the invention is a few years off at best, but serious drinkers better hope this thing automatically alerts you when the beans are running low, too.[Image courtesy of CoffeeToThePeople]
Hacker gets revenge on Puzzle Quest's bullying AI
We're not certain what sort of demonic sacrifices Infinite Interactive made to grant Puzzle Quest its hellborn AI, but we imagine that the cursed contract that authenticated the ceremony was written with the blood of many innocents. The match-three puzzler does everything short of outright cheating, stealing your advantages and setting up multiple combos, each computer-cleared gem bearing the mark of Mephistopheles.DS gamer Zaraf plotted a strategy that would tear down the AI's defenses and avenge dozens of unfair losses. Unwilling to spend months leveling up and making preparations, staying his vendetta, he hex-edited the game to to max out his character's stats. Zaraf then armed his warrior with a class spell called Deathbringer, enabling him to fill the screen with an amount of damaging skulls equal to half of his red mana. Head past the post break for the results caught on video.
Puzzle Quest's AI doesn't cheat, but you can!
If the number one complaint gamers have with Puzzle Quest is its limited availability at game shops, then the second most common point of protest would be the Puzzle/RPG's cheating AI. People are just as apt to sing praises about its addictive gameplay as they are to howl over the AI's godlike prescience. We've spent more than a few battles shaking our fists at the game as computer-controlled enemies racked up lucky combos and more extra turns than chicken on a rotisserie.Sensing that the mob was two forums threads away from storming his house with torches and pitchforks, Infinite Interactive's Steve Fawkner made a public statement assuring players that the AI has no unseen advantages. Having worked on the code himself, Steve reasoned that he's too lazy to have programmed anything that advanced.If that explanation isn't convincing enough, there are still steps you can take to even the playing field. You can unlock a debug menu by pushing in a complex set of keypresses, allowing you to activate several hidden features. Check past the post break for more details on the cheat code and a comic about Puzzle Quest's AI.
University of Essex developing autonomous model car
DARPA's Grand Challenge certainly snags a majority of the spotlight when talking about autonomous vehicles, but researchers at the University of Essex are looking to tackle the idea on a (literally) smaller scale. Seeking to craft a "driverless model car," the project will reportedly utilize a standard remote control model vehicle, which will be flanked by a PC, camcorder, and a bevy of sensors. Supposedly, the software that will be riding on board will allow the vehicle to be "entirely autonomous" by recognizing obstacles, making tactical decisions, and driving itself around a test track. The team responsible for the prototype hopes that this small-scale, low-cost endeavor will "pave the way" for autopilot cars of the future, and considering the problems we mere humans are already having with newfangled technology, that day can't come soon enough.[Via Slashdot]
SXSW: ARG! The Attack of the Alternate Reality Games
Ever since the success of The Beast, the alternate reality game created to help pimp Spielberg's A.I. back in 2001, alternate reality games (ARGs) have been popping up left and right, most notably the I Love Bees ARG that was used to launch Halo 2. Based on what the panelists were telling us, there are a lot more coming down the pipeline.However, one of the problems was that the panel promised to help define the term "alternate reality game," but that never happened. Wikipedia calls it "an interactive narrative that uses the real world as a platform, often involving multiple media and game elements, to tell a story that may be affected by participants' ideas or actions." Which is quite a mouthful. But that makes us wonder, does it have to use the web as a medium to be an ARG? When people used to play T.A.G. or Killer on college campuses, that was definitely an ARG ... but where did those games go?
AI program slammed for practicing law without a license
While artificial intelligence programs offering legal advice aren't exactly anything new, as Wired's 27B Stroke 6 reports, it looks like we've now seen the first case of one running into trouble with the law for doing so. The over-eager AI in question was offering its services to entrepreneur Henry Ihejirika, who put the program to use on two of his websites, offering bankruptcy assistance to clients without the hassle of a face-to-face meeting. Things were apparently going swimmingly until a bankruptcy trustee noticed errors in some of the forms that were submitted by a client of the site, which led them to investigate the situation, ultimately resulting in Iherjirka heading to court to explain himself. After reviewing the case, a bankruptcy judge ruled that the software went far beyond simply providing clerical services and was, in effect, practicing law without a license. That meant Ihejirka had to pull the plug on the system, as well as pay fines and return all fees he had collected from clients. While the AI could not be reached for comment, it'll no doubt find plenty of work on the inside, helping out prison guards with their taxes.[Via Boing Boing]
PaPeRo gets blogging software; Engadget one step closer to full automation
Now we were always under the impression that a personal blog is supposed to be just that -- personal -- so we're not sure that we see the advantages of bringing a robot in to automate this process by filling your site with multimedia content that it thinks you'd want to share with the rest of the world. Nonetheless, NEC has done exactly that with its little cannibalistic PaPeRo bot (you know, the one that thinks humans taste like bacon), endowing it with AI software that recognizes certain keywords uttered during a conversation with its master and then scours the net for seemingly-related pics, vids, and tunes. Scheduled to be unveiled at 13th Annual Conference of the Association of Natural Language Processing later this month in Japan, the newly-spec'ed PaPeRo will be tasked with listening to you talk about your boring day at work ("So I commuted the eight feet from bed to desk, blogged all morning, ate lunch, blogged some more, ate dinner, and then blogged until bedtime."), and then turning your page into what we can only imagine will be a blinking, flashing, slow-loading lookalike of some teen's gaudy MySpace. Just be careful what you talk about from now on, because PaPeRo may be listening, and the last thing you want on your blog is a visual representation of that thing you've been doing to your coworkers' coffee every morning for the last eight weeks.[Via Digital World Tokyo]
Watch out Stanley, here comes Junior
With the slow vehicle passing and a 50mph speed limit, the 2005 DARPA Grand Challenge didn't entirely seem to set the stage technologically for the 2007 DARPA Urban Challenge, but get ready, Stanford's already prepping, their entrant: Junior. The Volkswagen Passat wagon will be equipped with a 360-degree laser rangefinder, bumper mounted lasers, RADAR, GPS, a network of systems and software powered by Core 2 Duo processors, and hopefully also spinners to distract the competition's junk-ass rides. Junior's mission, if you choose to recall it: drive a simulated urban course 60 miles long; it must obey California state traffic laws, it must not crash, it must be able to operate without GPS, and it must run the course entirely without human input. The $2m at stake for first place is probably not nearly enough to immediately recoup the costs of a bunch of braniac grad students hacking complex AI algos, but it could be the icing on the cake for the current favorite after 2005's Grand Challenge was routed by Junior's pappy, Stanley.P.S. -CNET has some early pics of Junior's interior and such, check 'em out.
Sony Ericsson's Walkman W880 reviewed
Sony Ericsson's venerable Walkman line just received a good dose of cool with the announcement and (pending?) release of the W880. Sony Ericsson has joined the thin-is-in crowd with this one, as the W880 comes in at a skinny and svelte 9.4mm thick -- pretty dainty for a candybar unit. This is one of the newest Sony Ericsson handsets we've been pleased to view and gawk at. Slim, trim, and comes in fightin' with such features like 2 cams (a VGA one for front-facing video calls), Bluetooth 2.0 A2DP and Memory Stick M2 support (alas, Sony's proprietary format) and...we'll say it again -- an ultra-slim shape that's more Samsung-ian than from the SE boys. A few things that stood out in the GSM Arena review was the inclusion of "multitasking" support (we can guess at the meaning there), EDGE being absent and the inclusion of a standard 3.5mm audio adapter for using your own high-end headphones with the W880. We wish all handset makers would do this -- since using those sweet Shure earphones are, well, worth it. GSM Arena's verdict with the in-depth review? This is one sweet candybar.[Thanks, Simon]
Sony Ericsson's W880 (Ai) Walkman musicphone unleashed
Just like we thought, Sony Ericsson officially launched their W880 (Ai) Walkman musicphone today. Good thing too, 'cause all the leaked photos and SE teasers were getting a bit tiresome. So was the wait worth it? Depends, do you like slim musicphones swaddled in brushed stainless steel? At just 9.4-millimeters thin, it's SE's slimmest Walkman phone yet and packs a 1.8" QVGA 262k TFT display and 2 megapixel camera with a 1GB Memory Stick Micro (M2) card included in the box. It also comes pre-loaded with Walkman Player 2.0, Disc2Phone music management, and the TrackID music recognition applications. It also features a Flight Mode for use on the airplane and battery capable of up to 20-hours of music playback or up to 6 hours 30 mins of talk time. Good so far right? Sadly for those of us in The States, it'll only be sportin' UMTS 2100 and GSM 900/1800/1900 when it hits Europe in Q1. A GSM-only variant dubbed the W888 is headed to China. Check the gallery below for a taste of what might have been. %Gallery-1476%
