Big Brother is watching you!

[buglerbilly]

Darpa Wants Code to Spot ‘Anomalous Behavior’ on the Job

By Noah Shachtman May 20, 2010 | 10:22 am



Can software catch a cyberspy’s tricky intentions, before he’s started to help the other side? The way-out researchers at Darpa think so. They’re planning a new program, “Suspected Malicious Insider Threat Elimination” or SMITE, that’s supposed to “dynamically forecast” when a mole is about to strike. Also, the code is meant to flag “inadvertent” disclosures “by an already trusted person with access to sensitive information.”

“Looking for clues” that suggest a turncoat or accidental leaker is about to spill (.pdf) “could potentially be easier than recognizing explicit attacks,” Darpa notes in a request for information. But even that simpler search won’t be easy. “Many attacks are combinations of directly observable and inferred events.” Which is why SMITE’s program managers are interested in techniques to figure out “the likely intent of inferred actions, and suggestions about what [that] evidence might mean.” That goes for “behaviors both malicious and non-malicious.”

Step one in starting that process: Build a ginormous database to store all kinds of information on would-be threats. “The next step is to determine whether an individual or group of individuals is exhibiting anomalous behavior that is also malicious.” That’s a toughie — something anomalous in one context might be perfectly normal in another. One possible solution, the SMITE paper adds, could be detecting “deceptive” activities, which are a sign of cyberspying. Or cheating on your taxes. Or carrying on an office affair. Or playing World of Warcraft on the job. Depending on the situation.

Over at The Register, Lew Page quips: “It will no doubt be a comfort for anyone in a position of trust within the U.S. information infrastructure to know that mighty military algorithms and hybrid engines will soon sniff your every move so as to forecast any context-dependent malice on your part — and then in some unspecified way (remember what the E in SMITE stands for) eliminate you as a threat.”

More likely, the program is just a way to do some basic research into algorithms’ ability to understand human intent. But since every Darpa program has to have some sort of military application — no matter how far-fetched — the agency has cooked up this cyberspy-fighting scenario.

Anyway, our spies tell us that Darpa is planning a SMITE workshop for mid-June in northern Virginia.

Photo: Flickr/StuartPilbrow

Read More http://www.wired.com/dangerroom/2010/05/darpa-wants-code-to-spot-anomalous-behavior-on-the-job/#more-25041#ixzz0oWMQ2vb0

[buglerbilly]

Darpa’s Self-Learning Software Knows Who You Are

By Katie Drummond May 21, 2010 | 9:48 am



Software systems could one day analyze everything from blurry war-zone footage to the subtle sarcasm in a written paragraph, thanks to two unassuming scientists who are inspired by biology to make revolutionary strides in intelligent computing.

Yann LeCun and Rob Fergus, both computer science professors at New York University, are the brains behind “Deep Learning,” a program sponsored by Darpa, the Pentagon’s blue-sky research agency. The idea, ultimately, is to develop code that can teach itself to spot objects in a picture, actions in a video, or voices in a crowd. LeCun and Fergus have $2 million and four years to make it happen.

Existing software programs rely heavily on human assistance to identify objects. A user extracts key feature sets, like edge statistics (how many edges an object has, and where they are) and then feeds the data into a running algorithm, which uses the feature sets to recognize the visual input.

“People spend huge amounts of time building these feature sets, figuring out which are better or more accurate, and then refining them,” LeCun told Danger Room. “The question we’re asking is whether we can create computers that automatically learn feature sets from data. The brain can do it, so why not machines?”

The computer systems will be inspired by biology, but not modeled after it. That’s because researchers still aren’t entirely sure how animals are able to turn inputs — an object, a movement, a sound — into usable information. Ten years ago, a study at MIT helped answer the question. Researchers rewired ferret brains, so that the optical nerve fed into the auditory cortex, and vice versa. But the ferrets still saw and heard normally, leading the team to conclude that brain function depends on the signal — not the area.

Brains also display plenty of abstraction when it comes to identifying specific inputs: LeCun was inspired to create his algorithmic layering approach, called “a convolutional network,” by the 1960s research of David Hubel and Torstein Weisel. The two used cats to demonstrate how the brain’s visual cortex relies on abstractions to create complex representations of a given visual input.

In other words, LeCun said, “There’s some sort of learning algorithm within the brain. We just don’t know what it is.”



But the algorithmic talents of the mind, along with its ability to identify visual data by abstraction, will be the key components of the NYU team’s new system. Right now, an algorithm recognizes objects in one of two ways. In one, it is shown some representative examples of what, say, a horse looks like. Then the code tries to match any new creature to the ur-stallion. (That’s called “supervised” learning.) In the other way, the software is shown lots and lots of horses, and it builds its own model of what a horse is supposed to resemble. (That’s “unsupervised” learning.)

What LeCun and Fergus are trying to do is make code that can get it right on a first, unsupervised example — using layer after layer of code to abstract the essential attributes of an object. This first step is to turn an image into numbers: For a 100 x 100 pixel image, the software produces a grid of 10,000 numbers; 9 x 9 “masks” are then applied to that grid, to uncover attributes of the image. The first feature spotted is an object’s edge. (The human brain makes a similar initial pass.) Several more “masks” follow. The final output? A series of 256 numbers that identifies the input.

The two are only six weeks into the project, but they’ve already got demos up and running.

The Deep Learning algorithm and I had never met, but with a quick shot by a small webcam on LeCun’s laptop, the layers of code captured my features and could immediately distinguish me from other objects and people in LeCun’s office. The same thing happens when LeCun introduces the system to two different coffee mugs — it takes mere seconds for the computer to acquaint itself with each, then distinguish one from the other.

And this is only the beginning. Darpa also wants a system that can spot activities, like running, jumping or getting out of a car. The final version will operate unsupervised, by being programmed to hold itself accountable for errors — and then auto-correct them at each algorithmic layer.

It should also be able to apply the layered algorithmic technique to text. Right now, computer systems can parse sentences to categorize them as positive or negative, based on how often different words appear in the text. By applying layers of analysis, the Deep Learning machine will — LeCun and Fergus hope — spot sarcasm and irony too.

“Ideally, what we’ll come away with is a ‘generic learning box’ that can identify every data cue,” Fergus tells Danger Room.

Photo: Katie Drummond

Read More http://www.wired.com/dangerroom/2010/05/darpa-code-teaching-itself-what-the-world-looks-like/#more-25044#ixzz0obtoxs61

[buglerbilly]

Darpa’s Beady-Eyed Camera Spots the ‘Non-Cooperative’

By Katie Drummond May 27, 2010 | 4:12 pm



Soon, keeping your head down won’t be enough to stump high-tech security cameras, thanks to Pentagon-funded researchers developing mini-cameras that can nab threats by hunting down — and scanning — their eyeballs.

A team of electrical engineers at Southern Methodist University (SMU), led by Professor Marc Christensen, first created the cameras with funding from Darpa, the Pentagon’s research agency. Called Panoptes, the devices use low-resolution sensors to create a high-res image that can be captured using a lightweight, ultra-slim camera. Because they don’t use a lens, the cameras were originally designed for miniature drone sensors and troop helmet-cams.

Only a year later, the Pentagon is giving SMU another $1.6 million, to merge the cameras with active illumination and handheld Pico projection devices. This allows photos captured on small devices to be transformed for large-format viewing. Whereas the first goal of the program was to create slim cameras with the power of a lens, the latest technology “lets us do even more than what a lens could do,” Christensen told Danger Room.

“This platform is really just the base, upon which we’ll focus on different applications,” Christensen said. “Now, we’re enhancing resolution even more, so the images are a 3-D map with even better, more accurate details.”

The new devices will yield a robust 3-D image that’ll be useful for seeing in caves and dark urban areas, and for the creation of versatile “non-cooperative” iris-detection security cameras.

Smart-Iris, the name of the new Panoptes innovation, is being developed in conjunction with SMU Professor Delores Etter, who specializes in biometric identification. It’ll eliminate problems like glare, eyelashes, dim lighting — and an unwillingness to stop and stare directly into a dedicated iris-detection camera. Instead, Panoptes devices will zero in on a face, no matter angle or movement, then narrow right into the iris. A long line of people, moving through a line, could be scanned by wall-mounted cameras and they wouldn’t even notice it was happening.

And new algorithms are being developed by Etter and colleagues to identify individuals based on segments of their iris, rather than a full frontal scan.

“Ideally, when you walk down a hallway, no matter where your head is looking, the device can grab your eyeball and detect what it needs to,” Christensen said. And where possible security and defense applications are concerned? “You can let your imagination fly with that one.”

And with this latest development, Christensen also sees widespread civilian application, as part of “the cell phone of the future.” He’d like to see the camera-projection device incorporated into phones, and says they’d be able to photograph the page of a book “down to the smallest lettering,” or detect counterfeit cash by “picking up the texture of a $20 bill.”

Photo: U.S Air Force