ai

Robots

The Future is Coming to a Robot Near You (Or Behind, Rather)

Posted by Gizmodo US Edition at 10:45 AM on October 24, 2008

The Pentagon and the Department of Defence has recently put in a request to contractors for a multi-robot pursuit system to search and track down "non-cooperative humans." The military is worried that controlling robots will take time away from defence officials so creating a pack-hunting AI that will do it automatically will be useful. Once the system is perfected, government officials expect it to become autonomous and armed. Maybe next time, you'll think twice about littering or about to make a "California" stop. That's exactly what this economy needs anyways: a good use of taxpayers' dollars to replace more jobs. [New Scientist]

Robots

Computer Nearly Passes Turing Test for Artificial Intelligence

Posted by Jack Loftus at 6:00 AM on October 13, 2008

Today, the machines became a little smarter, as a computer named Elbot managed to achieve a 25% success rate when convincing a human being that they were talking to another human. The experiment is called the Turing Test, after mathematician Alan Turing, and Sunday's saw six Artificial Conversational Entities (ACEs) trying to ace the exam. Word is there was one human dunce in the mix, as all six computers managed to fool at least one interrogator into thinking they were speaking to another person, but none of the machines could officially pass Turing's strict standards.


Read More »

Vehicles

Stanford's UAV Helicopters Learn to Fly Themselves by Watching

Posted by Kit Eaton at 7:20 PM on September 2, 2008

They may not look as mean as the Draganfly, but these research UAV 'copters are one step closer to Skynet: they learn to fly complex stunts by just "watching" another aircraft do the same. Dubbed "apprenticeship learning," by the Stanford team which developed it, the system gets its flight plan by recording an expert human operator fly a vehicle. Then its onboard gyros and GPS systems and avionics communicate with a ground-based computer which looks at the human-derived data and decides how to fly the vehicle. It's all very tricky, due to the inherent instability of helicopter. Apparently UAVs like this may one day help firefighters track wildfires, or be sent to see out landmines in battlefield situations... and do other "spying" of course. Creepy. [Physorg]


Read More »

Science

Mars Phoenix Lander Protects Itself From Bad NASA Commands

Posted by Adrian Covert at 3:45 AM on July 17, 2008

The Mars Phoenix Lander shut its robotic arm down over the weekend, refusing to follow NASA directions after "realising" those actions would have damaged its wrist. NASA programmers had to send new code to bring the arm back to life, and are now augmenting the original code to try and get the task done. Seemingly pleased with the Phoenix's refusal to conform to The Man's rules, NASA representatives described the process as "pretty neat." I think this whole "machines thinking for themselves" thing is only neat until they decide all humans are off their collective asses, and leave them floating in space with no suit. [PC World via Slashdot]


Read More »

Robots

Poker Bot Knows When to Hold 'Em, Knows When to Fold 'Em

Posted by Gizmodo US Edition at 9:00 AM on July 13, 2008

Looks like the day when we all hang out with robots that smoke cigars, throw back drinks and win all our money in poker is edging closer, now that programmers at the University of Alberta have created a Hold 'Em-playing software that can beat poker professionals. The Polaris software was pitted against Poker pros like Nick Grudzien and Ijay Palansky in six games of limit Texas Hold 'Em--it tied on one, lost two and won three.


Read More »

Robots

New Yorker: Why We Won't Have Fully Conversational Robots

Posted by Adrian Covert at 9:50 AM on June 25, 2008

John Seabrook wrote a recent feature in The New Yorker about interactive-voice-response systems (I.V.R.) commonly used with customer service and tech support telephone hotlines. Seabrook spent time at B.B.N. Technologies watching these systems transcribe callers' words and analysing the tone of voice for emotions present. While breaking down the history of automated telephone services and voice recognition innovations, he attempts to tackle the larger question of whether or not we can create a fully conversational, quasi-conscious robot, akin to 2001: A Space Odyssey's Hal 9000. Judging from the number of experts interviewed for the piece, the answer is a resounding no.


Read More »

Software

Computer Writes 200,000 Books, Man Takes Credit

Posted by Mark Wilson at 12:30 AM on April 15, 2008

Using his self-designed AI program, Philip M. Parker has "authored" 200,000 books without ever lifting his pen. While his work features some self-explanatory medical texts like The Official Patient's Sourcebook on Acne Rosacea, most of Parker's publications are 100-300 page industry analysis pieces like DAGENHAM MOTORS GROUP PLC: Labour Productivity Benchmarks and International Gap Analysis. Parker uses his automated software to do basic market research (scouring the internet to pull various facts and figures), and then to completely assemble the book (with charts, a full table of contents and plenty of text). Each can take anywhere from 13 minutes to 3 hours. Here's a clip in which Parker shows off the process:


Read More »

Robots

Nexi, The Social Robot From MIT Goes For the Emo Look

Posted by Gizmodo US Edition at 12:50 AM on April 4, 2008

She may look miles away from crossing uncanny valley, but Nexi from MIT's Personal Robots Group is at least on the way. She's designed to be a "Mobile Social Dextrous" machine that moves like we do when we express emotions. So, she's got fully articulated arms and a head with features that can be motored around to form expressions. Acting out emotions, she's actually rather amazing, in a slightly sad robot kinda way: the video may send a few chills down your spine, no matter how "artificial" Nexi looks now.


Read More »

Science

iCub Baby Robot to Undergo Developmental Training, Still Isn't a Real Kid

Posted by Haroon Malik at 12:00 PM on March 2, 2008

Those spiffing fellows at the University of Plymouth, UK are undertaking a research project involving a baby-bot named iCub, which will see the robot actually learn how to speak. The three-foot high robot will help researchers deduce how language is taught, but the further reaching impact of the study include the prospect of developing humanoid robots that can learn, think and talk. Sure, Steven Speilberg has already envisioned such a future, but how far off was he?


Read More »

Gadgets

AI Tech Racers: Danger Dodging Slot Cars

Posted by Benny Goldman at 9:00 AM on February 21, 2008

These AI Tech Racers from Tomy are some of the more interesting slot-style racecars we've seen. The cars have built-in sensors to detect and avoid obstacles in their path—provided they have room to maneuver—and as a result, they can fly around a track without being locked into slots. Once started, the cars ride on their own, but you can give them a speed burst or hit the brakes if needed. Hit the jump to see these babies in action.


Read More »