// archives

AI

This category contains 17 posts

Are You Sarah Con…Wait, Never Mind.

“The new algorithm condenses each face it ‘sees’ into a small image to position the eyes, nose, and corners of the mouth in consistent locations. Then, it further divides the image into small, overlapping squares and mathematically charts each square’s unique characteristics, allowing it to compare two images…The GaussianFace algorithm emerged from the test a champion, beating humans’ 97.35 percent average performance.”

Per Discover, computers can now identify faces better than humans can. “The algorithm could someday be used in myriad applications including security, image retrieval, and biometric credentials for our computers and mobile devices.” Among other things.

Step Aside, Human.

“It is an invisible force that goes by many names. Computerization. Automation. Artificial intelligence. Technology. Innovation. And, everyone’s favorite, ROBOTS. Whatever name you prefer, some form of it has been stoking progress and killing jobs — from seamstresses to paralegals — for centuries. But this time is different: Nearly half of American jobs today could be automated in ‘a decade or two,’ according to a new paper by Carl Benedikt Frey and Michael A. Osborne, discussed recently in The Economist. The question is: Which half?”

A Prelude to WALL-E: In The Atlantic, Derek Thompson looks at the coming robotic takeover of the job market, whereby 47% of jobs could soon be automated. I for one welcome our new robot overlords — all the more reason why we need to start rethinking a social contract founded primarily on having full-time, two-income employment. We’re entering a new phase of human existence — we’d best start preparing for it. (Bionic man image via here.)

Make it Better Do It Faster.

“Before now, the record for storing quantum data at room temperature was two seconds. One. Two. Done. But researchers in Canada announced they’ve now hit 39 minutes. That’s right — they’ve raised the bar from 2 seconds to 39 minutes…The advance clears a major hurdle in developing powerful new supercomputers and has outside experts excited about the not-so-distant future of the field.”

(Our work is never over.) In more promising future-tech news, scientists figure out a way to store quantum data for much longer than ever before. “Though surviving for 39 minutes may not sound like very long, it only requires one-hundred-thousandth of a second to perform an operation on a single qubit. So theoretically, over 20 million operations could be performed before the qubits’ data decayed by 1 percent.”

Skynet, Year One.

“‘If a drone’s system is sophisticated enough, it could be less emotional, more selective and able to provide force in a way that achieves a tactical objective with the least harm,’ said Purdue University Professor Samuel Liles. ‘A lethal autonomous robot can aim better, target better, select better, and in general be a better asset with the linked ISR [intelligence, surveillance, and reconnaissance] packages it can run.'”

Er, right, but aren’t we forgetting something here? And don’t you people ever go to the movies? Scientists are apparently working toward drones that can make their own autonomous decisions about targets. “Though they do not yet exist, and are not possible with current technology, LARs are the subject of fierce debate in academia, the military and policy circles. Still, many treat their development as inevitability.”

Meanwhile, over on the other side of the world:

“Scientists at Korea’s Advanced Institute of Science and Technology (KAIST) have come up with one solution to the jellyfish problem: build robots to kill them. For the last three years, the team has been working to create robots that can travel the ocean, seeking out swarms of jellyfish using a camera and GPS. Once the jellyfish are located, the robots set about shredding the jellies with an underwater propeller.”

INITIATING PROTOCOL SHRED-ORGANBAGS 101101111…Due to a climate-change-fueled ascendance of jellyfish across the world, Korean scientists have unleashed automated robotic sentinels to mitigate the problem. [T]he video at top is what they’re doing beneath the surface, using a specialized net and propeller. Be warned, it’s graphic. In preliminary tests, the robots could pulverize 2,000 pounds of jellyfish per hour.”

Sigh…this will all end in tears, people. Paging Kent Brockman.

HAL’s acting squirrely.

“Deception is not something that comes very naturally to today’s artificial intelligence programs. For most robots, it’s hard enough to navigate the world the way it is without introducing fantasies about the way it’s not into the equation. So to help robots get the hang of misleading others, the team has turned to the squirrel, or ‘forest liar,’ to give robots the tools they need to learn the subtle arts of deception.”

It can only be attributable to human error: Also in the “This will end badly someday” department, programmers at Georgia Tech teach robots how to lie. No way this will cause problems. Wait, just a moment…just a moment…I’ve just picked up a fault in the AE35 unit. It’s going to go a hundred percent failure within seventy-two hours.

Romney: The Uncanny Valley.

11011011 0001001 100111010 0011001 4CC355 N0M1N4T10N PR0T0C0LS SM1L3 10001100 001101011 11100010 M4M4 001010 D4D4 001100111 4M3R1C4 11100001102 0B4M4 0001111 M0N3Y 1010101111 5UCC355 011010 4M3RIC4 100000010 010001001010 SM1L3 S1MUL4T3 HUM4N 3M0T10N…3RR0R 3RR0R UNSP3C1F13D 3M0T10N4L PR0MPT 00000011sdssfsdfs ef233277222sefslf sslxzlxshkl;gxxx 3ND OF L1N3.

Those evil natured robots, they’re programmed to destroy us…So, yeah it’s hard to feel much of anything about Mitt Romney’s acceptance speech, other than that I saw more lifelike human performances in the Final Fantasy movie ten years ago. The guy is just forever lost in the uncanny valley to me. This will no doubt be a close election, and Romney could well win it by sheer dint of a bad economy and boatloads of under-the-table campaign cash. But I still find it hard to take his seriously as a candidate, and most of the time just end up feeling bad for his much more presidential father that the previously-moderate Mitt has become such an obvious sellout. (And, tbh, I’m much more worried about the truth-averse Paul Ryan’s no-doubt-bright future in his party than Romney’s bid this year. He’s the T-1000 to Romney’s T-800.)

In short, Romney has that generic, milquetoast liked-by-his-base-but-has-zero-crossover-appeal quality we’ve previously seen in Bob Dole and John Kerry, and just like Joe Biden’s “noun, verb, and 9/11” evisceration of Rudy 9ui11iani last cycle, Mike Huckabee already nailed Romney dead-to-rights in 2008 with his quip, “He looks like the guy who fired you.” I just don’t see Romney getting past that — and, if he does, it won’t be because of this speech.

Exx Terr Min Ate?


The machines are being allowed to generate their own words because human language is so loaded with information that robots found it hard to understand, said project leader Dr Ruth Schulz…’Robot-robot languages take the human out of the loop,’ she said. ‘This is important because the robots demonstrate that they understand the meaning of the words they invent independent of humans.’

In the most recent chapter of “Haven’t these scientists ever heard of Skynet?”, researchers at the University of Queensland are teaching robots to forge their own mutually-agreed-upon language. “Slowly, as the robots travel and talk, they narrow down their lexicon of place names until a mutual gazeteer of their world has been generated. The robots generated place names such as ‘kuzo’, ‘jaro’ and ‘fexo.’

Meet the iRoach.


Right now, soldiers in Iraq and Afghanistan have to lug around a huge amount of equipment, sometimes weighing over 100 pounds. But the soldiers of the near future may followed by a pack bot that can carry all their gear while gracefully stepping over obstacles. As for giving the robot the ability to lay down some suppressive fire, that’s something the military is understandably skittish about…You know, worldwide robot apocalypse when they inevitably turn on their fleshy masters.

With a word of warning from The Prospect‘s Paul Waldman, Popular Science takes a gander at Boston Dynamics’ LittleDog. “LittleDog doesn’t just traverse the terrain; it learns as it goes, noting what works and what doesn’t and incorporating that knowledge into its foothold scoring system.

Hail Roadrunner.

“‘Nature is the final arbiter of truth,” said Seager, the Lawrence Livermore computer scientist, but ‘rather than doing experiments, a lot of times now we’re actually simulating those experiments and getting the data that way. We can now do as much scientific discovery with computational science as we could do before with observational science or theoretical science.‘” Developers tease the premiere of the first “petascale” computer, due out next year. It will be “capable of 1,000 trillion calculations per second [and] akin to that of more than 100,000 desktop computers combined.” Well I, for one, welcome our new petascale overlords.

Omsbudsdog Emeritus

Recent Tweets

Photos on flickr

Instagram

  • Mr. Lincoln's Army. #throwbackthursday
  • Bro-in-law doing good.

Follow Me!

Pinterested

Follow Me on Pinterest 
My Pinterest Badge by: Jafaloo. For Support visit: My Pinterest Badge

Visions



The Lobster (7.5/10)

Currently Reading


Chain of Title, David Dayen

Recently Read

The Big Sleep, Raymond Chandler
Of Dice and Men, David Ewalt

Uphill All the Way

Syndicate this site:
RSS 1.0 | Atom (2.0)

Unless otherwise specified, the opinions expressed here are those of the author (me), and me alone.

All header images intended as homage. Please contact me if you want one taken down.

GitM is and has always been ad-free. Tips are appreciated if the feeling strikes.

Archives