Tomorrow is Today

Self Driving Cars Process .75 GB of Data Every Second, Including Cigarettes and Bouncing Balls

Google’s self-driving car soaks in a staggering amount of information about its environment as it drives, gathering 750MB of data every second, claims Idealab founder and CEO Bill Gross.

Gross, who says he learned the fact and taken the image above at an XPrize event, says the car is capturing everything it sees while moving — even cigarette butts.

“If it sees a cigarette butt, it knows a person might be creeping out from between cars. If it sees a rolling ball it knows a child might run out from a driveway. I am truly stunned by how impressive an achievement this is,” wrote Gross in a blog post.

(via This Is What Google’s Self-Driving Car ‘Sees’ as It Makes a Turn)

The End of Psychology: Diagnosing Depression With Kinect

Computer scientists at the University of Southern California have used Microsoft’s Kinect sensor to detect, with 90% accuracy, whether you are depressed. All you have to do is sit down in front of Kinect, answer some questions from an on-screen virtual psychologist, and the clever software does the rest. The process is entirely automated, objective, and self-contained, meaning accurate testing could be carried out in complete privacy at home.

The software, called SimSensei and developed by Stefan Scherer and colleagues, is essentially a clever mix of computer vision algorithms and the psychological model of depression. The on-screen psychologist asks you leading questions — a lot like the old-school Eliza, or Alice — and then watches how you physically respond. Using Kinect, the computer vision algorithms build up a very detailed model of your face and body, including your “smile level,” horizontal gaze and vertical gaze, how wide open your eyes are, and whether you are leaning toward or away from the camera. From these markers, SimSensei can work out whether you’re exhibiting signs that indicate depression — gaze aversion, smiling less, and fidgeting. Watch the video below and be amazed.

(via Kinect-based system diagnoses depression with 90% accuracy | ExtremeTech)

Australian Computer Scientists Develop  Digital Face to Add Emotion to Synthesized Speech

How it works is, the user types a phrase for Zoe to say. Six sliders allow you to set the emotions; for example, you could combine happiness and anger, setting them to halfway or full strength, depending on what you want her to convey. Then you can slow down or speed up her speech, giving a pretty large array of tone.

When tested with a group of 20 volunteers, they were able to accurately guess the emotion 77 per cent of the time — more than with the real-life Zoë, with whom the success rate was 73 per cent.

The team sees Zoe being used in the future as a personal assistant, but there are other potential applications as well, because the framework for the face is very light — tens of MBs — which means that it can be incorporated into small devices. It could also enable people to upload their own faces and voices into the program; the team envisions these being used as sort of “face messages” rather than text messages.

(via Virtual talking head can express human emotions - Crave)

DARPA Funded Research Aims to Automate Video Surveillance, Teach Computers to Identify Suspicious Behavior

This approach relies heavily on advances by machine vision researchers, who have made remarkable strides in last few decades in recognizing stationary and moving objects and their properties.

It’s the same vein of work that led to Google’s self-driving cars, face recognition software used on Facebook and Picasa, and consumer electronics like Microsoft’s Kinect.

When it works well, machine vision can detect objects and people — call them nouns — that are on the other side of the camera’s lens. But to figure out what these nouns are doing, or are allowed to do, you need the computer science equivalent of verbs.

…that’s where Oltramari and Lebiere have built on the work of other Carnegie Mellon researchers to create what they call a “cognitive engine” that can understand the rules by which nouns and verbs are allowed to interact. Their cognitive engine incorporates research, called activity forecasting, …which tries to understand what humans will do by calculating which physical trajectories are most likely.

They say their software “models the effect of the physical environment on the choice of human actions.”

Both projects are components of Carnegie Mellon’s Mind’s Eye architecture, a DARPA-created project that aims to develop smart cameras for machine-based visual intelligence. Predicts Oltramari: “This work should support human operators and automatize video-surveillance, both in military and civil applications.”

(via U.S. looks to replace human surveillance with computers | Security & Privacy - CNET News)

Robots That Learn Through Demonstration Instead of Programming

Neil Dantam at the Georgia Institute of Technology in Atlanta and colleagues are developing a system in which robots learn by example. Instead of programming a robot to carry out a task, a demonstration could suffice.

However, the robot does not simply mimic the human’s actions. The prototype system uses an Xbox Kinect camera to observe the human performing an activity, then breaks that activity into a sequence of key actions necessary to carry out the task.

These actions are converted into a general set of instructions – much like those found in an IKEA furniture instruction manual – that can be interpreted even by non-humanoid robots. “You want to somehow capture the important aspects of the human’s motion and transfer that to the robot,” says Dantam. “Think about how you’d tell someone how to make a cake,” he says. The way people follow those steps may vary. “An adult might bend down over the counter to work while a child may stand on tiptoes.”

(via Robot learns using IKEA-style instructions - tech - 10 October 2012 - New Scientist)

AI Developed to Improve Relevance of Commentator Patter During Live Sports

WHEN watching sport on TV …a good commentator can make all the difference, peppering a play-by-play account of the action with expert knowledge and anecdotes. But even the best commentator’s repertoire is limited.

…Created by sports fan Greg Lee at the University of Alberta in Canada, [AI Program] Scores can tap into a stash of sporting stories to find relevant anecdotes that a commentator might not have thought of.

…The system works by matching the features of a live event - such as the teams, key players, the score and the remaining time - against a database of available stories. Once stories that include some of those features are found it selects the few that are most relevant and suggests them to a human commentator.

The challenge is in evaluating the relevance of candidate stories and ranking them. Lee’s system uses machine-learning techniques to do this. The most important feature was the teams involved and the second was the difference in number of runs.

To test the system, the researchers used it to create commentary for pre-recorded sports broadcasts and presented them to 254 volunteers, who said they found the commentary relevant and enjoyable. 

(via AI sports commentator knows all the best stories - tech - 04 October 2012 - New Scientist)

UPDATE: Google’s Cat-Video Identifying Neural Net Now Working to Improve Voice Commands, Image Search, Google Glass and Self-Driving Cars

Google is now using these neural networks to recognize speech more accurately, a technology increasingly important to Google’s smartphone operating system, Android, as well as the search app it makes available for Apple devices.

“We got between 20 and 25 percent improvement in terms of words that are wrong,” says Vincent Vanhoucke, a leader of Google’s speech-recognition efforts. “That means that many more people will have a perfect experience without errors.”

The neural net is so far only working on U.S. English, and Vanhoucke says similar improvements should be possible when it is introduced for other dialects and languages.

Other Google products will likely improve over time with help from the new learning software. The company’s image search tools, for example, could become better able to understand what’s in a photo without relying on surrounding text. And Google’s self-driving cars and mobile computer built into a pair of glasses could benefit from software better able to make sense of more real-world data.

(via Google Puts Its Virtual Brain Technology to Work - Technology Review)

See Also: “Google and Stanford have created the [digital equivalent of the] visual cortex of an infant human”

Algorithm Analyzing Artistic Style Identifies Same Patterns as Art Historians

Computer scientists Computer scientists… have developed a program that analyzes paintings in a manner similar to how expert art historians perform their analysis, and conducted an experiment that showed that machines can outperform untrained humans in the analysis of fine art.

In the experiment, the researchers used approximately 1, 000 paintings of 34 well-known artists, and let the computer algorithm analyze the similarity between them based solely on the visual content of the paintings, and without any human guidance.

Surprisingly, the computer provided a network of similarities between painters that is largely in agreement with the perception of art historians.

I take this research with a grain of salt: the algorithm seems to have detected the patterns it was designed to detect.  I would have been more excited if it drew a new connection that human art experts found interesting.

(via Can computers understand art? | KurzweilAI)

Italian Roboticist Wants to Study Consciousness By Teaching Japanese Robots to Sing Jazz

Intelligence is often defined as the ability to find connections between existing entities - understanding that a key goes in a lock, for instance.

But Chella suggests that a conscious organism should be able to go a step further and introduce novel connections - between, say, musical phrases - that result in the creation of something new. That, in essence, is the idea behind improvisation…

Chella wants to replicate these states in a machine. “Consciousness could be linked to these moments of combination,” he says.

“[This work] raises interesting questions about the link between consciousness and music making,” says musician and computer scientist Philippe Pasquier… But he is sceptical about whether a robot musician needs a physical body, citing examples of AI composers that exist only in software.

(via Jazz-singing robot could shed light on consciousness - tech - 27 September 2012 - New Scientist)

Military Reviewing “Optionally Manned” Aircraft to Bridge Gap Between Current, Future Platforms

Today’s technology is insufficient to allow unmanned aircraft to make independent, complex judgments in an ambiguous and novel environment and so must be tethered to human judgment. But data links can be denied or deceived, and in any case, they introduce seconds of delay that can be crucial, such as when prosecuting moving targets or surviving in well-defended airspace.

In theory, some of these limitations ought to erode as unmanned technology develops. Aircraft can already take off and land by themselves. Funded programs are developing autonomous air refueling and “sense and avoid” systems, which will allow more regular flight operations in regulated airspace.

The Air Force is also funding research into tougher problems, such as automated dynamic mission planning and automated target recognition, and will continue to do so…

Academic centers also continue research into artificially intelligent moral reasoning that might one day allow autonomous application of rules of engagement in complex situations.

Finally, the history of electronics suggests that after 2020, computational power roughly equal to that of a human brain may be available to consumers for about what we pay today for a decent laptop.

Yet the technology that would enable true autonomy for a combat aircraft remains unusually hard and elusive. There is little reason to believe key technologies such as automated mission planning and target recognition will be ready in a decade, which is about how long it takes to design, build and test a major new aircraft, such as the LRS-B.

On the other hand, such an aircraft is likely to fly 20 to 40 years beyond its initial acquisition, a span that may see the emergence of key pacing technologies and perhaps even true human-equivalent general artificial intelligence.

The best option is to build future platforms “autonomy ready” — that is, so local or remote pilots could be replaced by artificial intelligence through software upgrades rather than costly hardware retrofits or new platforms.

Such design for optional manning would create an aircraft able to execute all aspects of its mission at initial operating capability while accommodating the overall trend toward greater unmanned capabilities. It neither assumes that key unmanned capabilities will be ready along with the airframe and propulsion systems nor locks out the kinds of autonomic improvements likely in the following decades.

(via The case for optionally manned aircraft | KurzweilAI)