A Semantic Sommelier: Wine Application Highlights the Power of Web 3.0

Web scientist and Rensselaer Polytechnic Institute Tetherless World Research Constellation Professor Deborah McGuinness has been developing a family of applications for the most tech-savvy wine connoisseurs since her days as a graduate student in the 1980s -- before what we now know as the World Wide Web had even been envisioned.

Today, McGuinness is among the world's foremost experts in Web ontology languages. These languages are used to encode meanings in a language that computers can understand. The most recent version of her wine application serves as an exceptional example of what the future of the World Wide Web, often called Web 3.0, might in fact look like. It is also an exceptional tool for teaching future Web Scientists about ontologies.

"The wine agent came about because I had to demonstrate the new technology that I was developing," McGuinness said."I had sophisticated applications that used cutting-edge artificial intelligence technology in domains, such as telecommunications equipment, that were difficult for anyone other than well-trained engineers to understand." McGuinness took the technology into the domain of wines and foods to create a program that she uses as a semantic tutorial, an"Ontologies 101" as she calls it. And students throughout the years have done many things with the wine agent including, most recently, experimentation with social media and mobile phone applications.

Today, the semantic sommelier is set to provide even the most novice of foodies some exciting new tools to expand their wine knowledge and food-pairing abilities on everything from their home PC to their smart phone. Evan Patton, a graduate student in computer science at Rensselaer, is the most recent student to tinker with the wine agent and is working with McGuinness to bring it into the mobile space on both the iPhone and Droid platforms.

The agent uses the Web Ontology Language (OWL), the formal language for the Semantic Web. Like the English language, which uses an agreed upon alphabet to form words and sentences that all English-speaking people can recognize, OWL uses a formalized set of symbols to create a code or language that a wide variety of applications can"read." This allows your computer to operate more efficiently and more intelligently with your cell phone or your Facebook page, or any other webpage or web-enabled device. These semantics also allow for an entirely new generation in smart search technologies.

Thanks to its semantic technology, the sommelier is input with basic background knowledge about wine and food. For wine, that includes its body, color (red versus white or blush), sweetness, and flavor. For food, this includes the course (e.g. appetizer versus entrée), ingredient type (e.g. fish versus meat), and its heat (mild versus spicy). The semantic technologies beneath the application then encode that knowledge and apply reasoning to search and share that information. This semantic functionality can now be exploited for a variety of culinary purposes, all of which McGuinness, a personal lover of fine wines, and Patton are working together on.

Having a spicy fish dish for dinner? Search within the system and it will arrive at a good wine pairing for the meal. Beyond basic pairings, the application has strong possibilities for use in individual restaurants, according to McGuinness, who envisions teaming up with restaurant owners to input their specific menus and wine lists. Thus, a diner could check menus and wine holdings before going out for dinner or they could enter a restaurant, pull out their smart phone, and instantly know what is in the wine cellar and goes best with that chef's clams casino. Beyond pairings, diners could rate different wines, providing fellow diners with personal reviews and the restaurateur with valuable information on what to stock up on next week. Is it a dry restaurant? The application could also be loaded up with the inventory within the liquor store down the street.

Beyond the table, the application can also be used to make personal wine suggestions and virtual wine cellars that you could share with your friends via Facebook or other social media platforms. It could also be used to manage a personal wine cellar, providing information on what is a peak flavor at the moment or what in your cellar would go best with your famous steak au poivre.

"Today we have 10 gadgets with us at any given time," McGuinness said."We live and breathe social media. With semantic technologies, we can offload more of the searching and reasoning required to locate and share information to the computer while still maintaining personal control over our information and how we use it. We also increase the ability of our technologies to interact with each other and decrease the need for as many gadgets or as many interactions with them since the applications do more work for us."

Source

Scientists Steer Car With the Power of Thought

They then succeeded in developing an interface to connect the sensors to their otherwise purely computer-controlled vehicle, so that it can now be"controlled" via thoughts. Driving by thought control was tested on the site of the former Tempelhof Airport.

The scientists from Freie Universität first used the sensors for measuring brain waves in such a way that a person can move a virtual cube in different directions with the power of his or her thoughts. The test subject thinks of four situations that are associated with driving, for example,"turn left" or"accelerate." In this way the person trained the computer to interpret bioelectrical wave patterns emitted from his or her brain and to link them to a command that could later be used to control the car. The computer scientists connected the measuring device with the steering, accelerator, and brakes of a computer-controlled vehicle, which made it possible for the subject to influence the movement of the car just using his or her thoughts.

"In our test runs, a driver equipped with EEG sensors was able to control the car with no problem -- there was only a slight delay between the envisaged commands and the response of the car," said Prof. Raúl Rojas, who heads the AutoNOMOS project at Freie Universität Berlin. In a second test version, the car drove largely automatically, but via the EEG sensors the driver was able to determine the direction at intersections.

The AutoNOMOS Project at Freie Universität Berlin is studying the technology for the autonomous vehicles of the future. With the EEG experiments they investigate hybrid control approaches, i.e., those in which people work with machines.

The computer scientists have made a short film about their research, which is available at:http://tinyurl.com/BrainDriver

Source

Augmented Reality System for Learning Chess

An ordinary webcam, a chess board, a set of 32 pieces and custom software are the key elements in the final degree project of the telecommunications engineering students Ivan Paquico and Cristina Palmero, from the UPC-Barcelona Tech's Terrassa School of Engineering (EET). The project, for which the students were awarded a distinction, was directed by the professor Jordi Voltas and completed during an international mobility placement in Finland.

The system created by Ivan Paquico, the 2001 Spanish Internet chess champion, and Cristina Palmero, a keen player and federation member, is a didactic tool that will help chess clubs and associations to teach the game and make it more appealing, particularly to younger players.

The system combines augmented reality, computer vision and artificial intelligence, and the only equipment required is a high-definition home webcam, the Augmented Reality Chess software, a standard board and pieces, and a set of cardboard markers the same size as the squares on the board, each marked with the first letter of the corresponding piece: R for the king (reiin Catalan), D for the queen (dama), T for the rooks (torres), A for the bishops (alfils), C for the knights (cavalls) and P for the pawns (peons).

Learning chess with virtual pieces

To use the system, learners play with an ordinary chess board but move the cardboard markers instead of standard pieces. The table is lit from above and the webcam focuses on the board, and every time the player moves one of the markers the system recognises the piece and reproduces the move in 3D on the computer screen, creating a virtual representation of the game.

For example, if the learner moves the marker P (pawn), the corresponding piece will be displayed on the screen in 3D, with all of the possible moves indicated. This is a simple and attractive way of showing novices the permitted movements of each piece, making the system particularly suitable for children learning the basics of this board game.

Making chess accessible to all

The learning tool also incorporates a move-tracking program called Chess Recognition: from the images captured by the webcam, the system instantly recognises and analyses every movement of every piece and can act as a referee, identify illegal moves and provide the players with an audible description of the game status. According to Ivan Paquico and Cristina Palmero, this feature could be very useful for players with visual impairment -- who have their own federation and, until now, have had to play with specially adapted boards and pieces -- and for clubs and federations, tournament organisers and enthusiasts of all levels.

The Chess Recognition program saves whole games so that they can be shared, broadcast online and viewed on demand, and can generate a complete user history for analysing the evolution of a player's game. The program also creates an automatic copy of the scoresheet (the official record of each game) for players to view or print.

The technology for playing chess and recording games online has been available for a number of years, but until now players needed sophisticated equipment including pieces with integrated chips and a special electronic board with a USB connection. The standard retail cost of this equipment is between 400 and 500 euros.

Source

Scientists Develop Control System to Allow Spacecraft to Think for Themselves

Professor Sandor Veres and his team of engineers have developed an artificially intelligent control system called 'sysbrain'.

Using natural language programming (NLP), the software agents can read special English language technical documents on control methods. This gives the vehicles advanced guidance, navigation and feedback capabilities to stop them crashing into other objects and the ability to adapt during missions, identify problems, carry out repairs and make their own decisions about how best to carry out a task.

Professor Veres, who is leading the EPSRC-funded project, says:"This is the world's first publishing system of technical knowledge for machines and opens the way for engineers to publish control instructions to machines directly. As well as spacecrafts and satellites, this innovative technology is transferable to other types of autonomous vehicles, such as autonomous underwater, ground and aerial vehicles."

To test the control systems that could be applied in a space environment, Professor Veres and his team constructed a unique test facility and a fleet of satellite models, which are controlled by the sysbrain cognitive agent control system.

The 'Autonomous Systems Testbed' consists of a glass covered precision level table, surrounded by a metal framework, which is used to mount overhead visual markers, observation cameras and isolation curtains to prevent any external light sources interfering with experimentation. Visual navigation is performed using onboard cameras to observe the overhead marker system located above the test area. This replicates how spacecraft would use points in the solar system to determine their orientation.

The perfectly-balanced model satellites, which rotate around a pivot point with mechanical properties similar to real satellites, are placed on the table and glide across it on roller bearings almost without friction to mimic the zero-gravity properties of space. Each model has eight propellers to control movement, a set of inertia sensors and additional cameras to be 'spatially aware' and to 'see' each other. The model's skeletal robot frame also allows various forms of hardware to be fitted and experimented with.

Professor Veres adds:"We have invented sysbrains to control intelligent machines. Sysbrain is a special breed of software agents with unique features such as natural language programming to create them, human-like reasoning, and most importantly they can read special English language documents in 'system English' or 'sEnglish'. Human authors of sEnglish documents can put them on the web as publications and sysbrain can read them to enhance their physical and problem solving skills. This allows engineers to write technical papers directly for sysbrain that control the machines."

Further information is available athttp://www.sesnet.soton.ac.uk/people/smv/avs_lab/index.htm.

Source

Future Surgeons May Use Robotic Nurse, 'Gesture Recognition'

Both the hand-gesture recognition and robotic nurse innovations might help to reduce the length of surgeries and the potential for infection, said Juan Pablo Wachs, an assistant professor of industrial engineering at Purdue University.

The"vision-based hand gesture recognition" technology could have other applications, including the coordination of emergency response activities during disasters.

"It's a concept Tom Cruise demonstrated vividly in the film 'Minority Report,'" Wachs said.

Surgeons routinely need to review medical images and records during surgery, but stepping away from the operating table and touching a keyboard and mouse can delay the surgery and increase the risk of spreading infection-causing bacteria.

The new approach is a system that uses a camera and specialized algorithms to recognize hand gestures as commands to instruct a computer or robot.

At the same time, a robotic scrub nurse represents a potential new tool that might improve operating-room efficiency, Wachs said.

Findings from the research will be detailed in a paper appearing in the February issue of Communications of the ACM, the flagship publication of the Association for Computing Machinery. The paper was written by researchers at Purdue, the Naval Postgraduate School in Monterey, Calif., and Ben-Gurion University of the Negev, Israel.

Research into hand-gesture recognition began several years ago in work led by the Washington Hospital Center and Ben-Gurion University, where Wachs was a research fellow and doctoral student, respectively.

He is now working to extend the system's capabilities in research with Purdue's School of Veterinary Medicine and the Department of Speech, Language, and Hearing Sciences.

"One challenge will be to develop the proper shapes of hand poses and the proper hand trajectory movements to reflect and express certain medical functions," Wachs said."You want to use intuitive and natural gestures for the surgeon, to express medical image navigation activities, but you also need to consider cultural and physical differences between surgeons. They may have different preferences regarding what gestures they may want to use."

Other challenges include providing computers with the ability to understand the context in which gestures are made and to discriminate between intended gestures versus unintended gestures.

"Say the surgeon starts talking to another person in the operating room and makes conversational gestures," Wachs said."You don't want the robot handing the surgeon a hemostat."

A scrub nurse assists the surgeon and hands the proper surgical instruments to the doctor when needed.

"While it will be very difficult using a robot to achieve the same level of performance as an experienced nurse who has been working with the same surgeon for years, often scrub nurses have had very limited experience with a particular surgeon, maximizing the chances for misunderstandings, delays and sometimes mistakes in the operating room," Wachs said."In that case, a robotic scrub nurse could be better."

The Purdue researcher has developed a prototype robotic scrub nurse, in work with faculty in the university's School of Veterinary Medicine.

Researchers at other institutions developing robotic scrub nurses have focused on voice recognition. However, little work has been done in the area of gesture recognition, Wachs said.

"Another big difference between our focus and the others is that we are also working on prediction, to anticipate what images the surgeon will need to see next and what instruments will be needed," he said.

Wachs is developing advanced algorithms that isolate the hands and apply"anthropometry," or predicting the position of the hands based on knowledge of where the surgeon's head is. The tracking is achieved through a camera mounted over the screen used for visualization of images.

"Another contribution is that by tracking a surgical instrument inside the patient's body, we can predict the most likely area that the surgeon may want to inspect using the electronic image medical record, and therefore saving browsing time between the images," Wachs said."This is done using a different sensor mounted over the surgical lights."

The hand-gesture recognition system uses a new type of camera developed by Microsoft, called Kinect, which senses three-dimensional space. The camera is found in new consumer electronics games that can track a person's hands without the use of a wand.

"You just step into the operating room, and automatically your body is mapped in 3-D," he said.

Accuracy and gesture-recognition speed depend on advanced software algorithms.

"Even if you have the best camera, you have to know how to program the camera, how to use the images," Wachs said."Otherwise, the system will work very slowly."

The research paper defines a set of requirements, including recommendations that the system should:

• Use a small vocabulary of simple, easily recognizable gestures.
• Not require the user to wear special virtual reality gloves or certain types of clothing.
• Be as low-cost as possible.
• Be responsive and able to keep up with the speed of a surgeon's hand gestures.
• Let the user know whether it understands the hand gestures by providing feedback, perhaps just a simple"OK."
• Use gestures that are easy for surgeons to learn, remember and carry out with little physical exertion.
• Be highly accurate in recognizing hand gestures.
• Use intuitive gestures, such as two fingers held apart to mimic a pair of scissors.
• Be able to disregard unintended gestures by the surgeon, perhaps made in conversation with colleagues in the operating room.
• Be able to quickly configure itself to work properly in different operating rooms, under various lighting conditions and other criteria.

"Eventually we also want to integrate voice recognition, but the biggest challenges are in gesture recognition," Wachs said."Much is already known about voice recognition."

The work is funded by the U.S. Agency for Healthcare Research and Quality.

Source

New Mathematical Model of Information Processing in the Brain Accurately Predicts Some of the Peculiarities of Human Vision

At the Society of Photo-Optical Instrumentation Engineers' Human Vision and Electronic Imaging conference on Jan. 27, Ruth Rosenholtz, a principal research scientist in the Department of Brain and Cognitive Sciences, presented a new mathematical model of how the brain does that summarizing. The model accurately predicts the visual system's failure on certain types of image-processing tasks, a good indication that it captures some aspect of human cognition.

Most models of human object recognition assume that the first thing the brain does with a retinal image is identify edges -- boundaries between regions with different light-reflective properties -- and sort them according to alignment: horizontal, vertical and diagonal. Then, the story goes, the brain starts assembling these features into primitive shapes, registering, for instance, that in some part of the visual field, a horizontal feature appears above a vertical feature, or two diagonals cross each other. From these primitive shapes, it builds up more complex shapes -- four L's with different orientations, for instance, would make a square -- and so on, until it's constructed shapes that it can identify as features of known objects.

While this might be a good model of what happens at the center of the visual field, Rosenholtz argues, it's probably less applicable to the periphery, where human object discrimination is notoriously weak. In a series of papers in the last few years, Rosenholtz has proposed that cognitive scientists instead think of the brain as collecting statistics on the features in different patches of the visual field.

Patchy impressions

On Rosenholtz's model, the patches described by the statistics get larger the farther they are from the center. This corresponds with a loss of information, in the same sense that, say, the average income for a city is less informative than the average income for every household in the city. At the center of the visual field, the patches might be so small that the statistics amount to the same thing as descriptions of individual features: A 100-percent concentration of horizontal features could indicate a single horizontal feature. So Rosenholtz's model would converge with the standard model.

But at the edges of the visual field, the models come apart. A large patch whose statistics are, say, 50 percent horizontal features and 50 percent vertical could contain an array of a dozen plus signs, or an assortment of vertical and horizontal lines, or a grid of boxes.

In fact, Rosenholtz's model includes statistics on much more than just orientation of features: There are also measures of things like feature size, brightness and color, and averages of other features -- about 1,000 numbers in all. But in computer simulations, storing even 1,000 statistics for every patch of the visual field requires only one-90th as many virtual neurons as storing visual features themselves, suggesting that statistical summary could be the type of space-saving technique the brain would want to exploit.

Rosenholtz's model grew out of her investigation of a phenomenon called visual crowding. If you were to concentrate your gaze on a point at the center of a mostly blank sheet of paper, you might be able to identify a solitary A at the left edge of the page. But you would fail to identify an identical A at the right edge, the same distance from the center, if instead of standing on its own it were in the center of the word"BOARD."

Rosenholtz's approach explains this disparity: The statistics of the lone A are specific enough to A's that the brain can infer the letter's shape; but the statistics of the corresponding patch on the other side of the visual field also factor in the features of the B, O, R and D, resulting in aggregate values that don't identify any of the letters clearly.

Road test

Rosenholtz's group has also conducted a series of experiments with human subjects designed to test the validity of the model. Subjects might, for instance, be asked to search for a target object -- like the letter O -- amid a sea of"distractors" -- say, a jumble of other letters. A patch of the visual field that contains 11 Q's and one O would have very similar statistics to one that contains a dozen Q's. But it would have much different statistics than a patch that contained a dozen plus signs. In experiments, the degree of difference between the statistics of different patches is an extremely good predictor of how quickly subjects can find a target object: It's much easier to find an O among plus signs than it is to find it amid Q's.

Rosenholtz, who has a joint appointment to the Computer Science and Artificial Intelligence Laboratory, is also interested in the implications of her work for data visualization, an active research area in its own right. For instance, designing subway maps with an eye to maximizing the differences between the summary statistics of different regions could make them easier for rushing commuters to take in at a glance.

In vision science,"there's long been this notion that somehow what the periphery is for is texture," says Denis Pelli, a professor of psychology and neural science at New York University. Rosenholtz's work, he says,"is turning it into real calculations rather than just a side comment." Pelli points out that the brain probably doesn't track exactly the 1,000-odd statistics that Rosenholtz has used, and indeed, Rosenholtz says that she simply adopted a group of statistics commonly used to describe visual data in computer vision research. But Pelli also adds that visual experiments like the ones that Rosenholtz is performing are the right way to narrow down the list to"the ones that really matter."

Source

Physicists Challenge Classical World With Quantum-Mechanical Implementation of 'Shell Game'

In a paper published in the Jan. 30 issue of the journalNature Physics, UCSB researchers show the first demonstration of the coherent control of a multi-resonator architecture. This topic has been a holy grail among physicists studying photons at the quantum-mechanical level for more than a decade.

The UCSB researchers are Matteo Mariantoni, postdoctoral fellow in the Department of Physics; Haohua Wang, postdoctoral fellow in physics; John Martinis, professor of physics; and Andrew Cleland, professor of physics.

According to the paper, the"shell man," the researcher, makes use of two superconducting quantum bits (qubits) to move the photons -- particles of light -- between the resonators. The qubits -- the quantum-mechanical equivalent of the classical bits used in a common PC -- are studied at UCSB for the development of a quantum super computer. They constitute one of the key elements for playing the photon shell game.

"This is an important milestone toward the realization of a large-scale quantum register," said Mariantoni."It opens up an entirely new dimension in the realm of on-chip microwave photonics and quantum-optics in general."

The researchers fabricated a chip where three resonators of a few millimeters in length are coupled to two qubits."The architecture studied in this work resembles a quantum railroad," said Mariantoni."Two quantum stations -- two of the three resonators -- are interconnected through the third resonator which acts as a quantum bus. The qubits control the traffic and allow the shuffling of photons among the resonators."

In a related experiment, the researchers played a more complex game that was inspired by an ancient mathematical puzzle developed in an Indian temple called the Towers of Hanoi, according to legend.

The Towers of Hanoi puzzle consists of three posts and a pile of disks of different diameter, which can slide onto any post. The puzzle starts with the disks in a stack in ascending order of size on one post, with the smallest disk at the top. The aim of the puzzle is to move the entire stack to another post, with only one disk being moved at a time, and with no disk being placed on top of a smaller disk.

In the quantum-mechanical version of the Towers of Hanoi, the three posts are represented by the resonators and the disks by quanta of light with different energy."This game demonstrates that a truly Bosonic excitation can be shuffled among resonators -- an interesting example of the quantum-mechanical nature of light," said Mariantoni.

Mariantoni was supported in this work by an Elings Prize Fellowship in Experimental Science from UCSB's California NanoSystems Institute.

Source