Carnegie Mellon Ties Knot (Again) with GM to Build
Autonomous Cars
Network World (06/19/08)
General Motors and Carnegie Mellon University recently announced that they
would fund and build a $5 million lab dedicated to developing autonomous
driving technologies. Carnegie Mellon and GM have a long history of
collaborating on driverless technology, including Carnegie Mellon Tartan
Racing's victory at the DARPA Urban Challenge last Fall with a GM SUV. GM
has also spent more than $11 million since 2000 on similar
Carnegie-partnered Collaborative Research Labs (CRL). The new CRL will
focus on developing a variety of autonomous technologies, including
electronics, controls, software, wireless capabilities, digital mapping,
and robotics. The lab will be located at Carnegie Mellon in Pittsburgh and
operate as an extension of GM's Global Research & Development network.
Many of the proposed technologies have already seen significant
advancement: For example, effective navigation and collision avoidance
technologies were standard in the DARPA Urban Challenge, where vehicles
constantly monitored the road ahead for vehicles and obstacles to avoid
collisions. Other autonomous car projects are also in progress. Using
technology created for the DARPA race, MIT AgeLab is working on its
AwareCar, which is a black Volvo with minicameras and infrared lights above
the steering wheel that monitor the driver's eye and eyelid movements.
Other sensors monitor the driver's heart rate, blood pressure, and
respiration to look for changes that the car should react to, like stopping
the car if the driver experiences heart pain. The AwareCar also has
monitors in the trunk to watch for lane drifting.
Click Here to View Full Article
to the top
Diamonds Offer Cool Computer Solution
ABC Science Online (Australia) (06/20/08) Salleh, Anna
University of Melbourne physicist Steven Prawer says the current
generation of computers are power hungry and inefficient, but that quantum
computers made using diamonds are a practical way to achieve a significant
improvement in computer power without generating more heat. Prawer says
quantum computers provide a new paradigm for computing that utilizes
exponential processing power through a highly efficient process that does
not create heat. Quantum computers will use "qubits" that can be on, off,
or both states at the same time depending on the electrons' spin, providing
extremely high processing power because messages based on different states
can be processed in parallel. Prawer says many quantum computer designs
rely on very low temperatures and complex infrastructures to detect the
electron spin and protect from being influenced by the outside environment,
but diamonds can provide a unique platform for building quantum computers
that can operate at room temperature. "All of the things that you would
want from a quantum computer have been demonstrated in diamond," says
Prawer. Tiny manufactured diamonds with a nitrogen atom at their center
can act as a qubit, and the spin of the electrons in the diamond can be
manipulated using microwaves or laser pulses. Although true quantum
computing is still years away, Prawer says diamonds can already be used for
a variety of new engineering and research devices, and that the first
quantum device to be commercialized was a diamond-based single photon
source used for quantum cryptography.
Click Here to View Full Article
to the top
Chill Out, Your Computer Knows What's Best for You
ICT Results (06/18/08)
Computers are continually becoming more user-friendly and increasingly
capable of anticipating the user's needs and acting to meet them, largely
because of the European Union-funded Computers in the Human Interactive
Loop (CHIL) project. The technologies developed by CHIL leave humans free
to concentrate on their objectives instead of having to think about the
computer and how to run the machine. The CHIL researchers examined ways
computers can serve humans better rather than forcing humans to adapt to
how computers are designed, and through their focus on human-machine
interaction they aimed to create a new paradigm of machine-supported
human-to-human interaction. The CHIL team developed systems that could
understand the context of meetings and proactively help participants by
controlling the meeting environment. Project scientific coordinator Rainer
Stiefelhagen says the project made "remarkable" achievements, highlighting
the project's advancements in building a new system of audio-visual
components to monitor and analyze what people are doing and how they behave
in different circumstances. One spin-off project that started two months
after CHIL ended will look at how police or fire officials in a crisis
management room handle incoming data during an emergency situation,while
another proposed project will assign a CHIL partner to a building company
to develop a smart house.
Click Here to View Full Article
to the top
Intel's Future Vision: Cars With Eyes, Processors With
Engines
Computerworld (06/12/08) Gaudin, Sharon
About 70 research projects that could impact the future were on display
during Research at Intel Day at the Computer History Museum in Mountain
View, Calif. The demonstration included using electronic fields to enable
machines to sense what is around them, a development that would allow a
robot to determine the amount of pressure to use in its fingers when
picking something up. With such a sense of touch, a robot would be able to
pick up glasses without breaking them or help a senior get up from the
couch. Intel is also building tiny specialized core engines that perform a
specific function, such as encryption or video acceleration. "If you can
build a chip and add a tiny, tiny engine inside the processor, then we can
actually use a lot less power, save electricity and lengthen battery life,"
says Manny Vara, the company's technology strategist. Intel showed off a
system with cameras and multicore processor-based computers that would
enable cars to determine when they are getting too close to pedestrians or
other vehicles and prompt a safety precaution. The company is also
optimistic about the use of speech recognition and interfaces to improve
the input capabilities of mobile Internet devices.
Click Here to View Full Article
to the top
IBM Launches Green Supercomputer in Sweden
InternetNews.com (06/17/08) Patrizio, Andy
IBM recently announced that it and Umea University in Sweden have
installed the most powerful Windows-based supercomputer in Europe. The
Akka supercomputer's performance is generally unremarkable, except for two
interesting aspects: Akka can run Windows and it is extremely green, using
about 40 percent less energy than a regular supercomputer of the same size,
according to Andreas Ryden, Nordic sales manager for IBM's High Performance
Computing market. The system is based on 672 IBM HS21XM blades that run
Intel's 2.5 GHz L5420 Xeon processor, which uses only 50 watts. The mix of
processors used in the supercomputer also allows researchers to use other
technologies, says Ryden. The computer will be used by researchers from
all over Sweden for projects including space science, material science,
bioinformatics, theoretical physics and chemistry, engineering science,
basic research in parallel algorithms, library software, and middleware for
grid infrastructure.
Click Here to View Full Article
to the top
BGU Researchers Develop New Gesture Interface
Device
Ben-Gurion University of the Negev (Israel) (06/18/08)
Ben-Gurion University of the Negev (BGU) researchers have developed a new
hand gesture recognition system that allows doctors to manipulate digital
images during medical procedures using motion, eliminating the need to
touch a screen, keyboard, or mouse, which compromises sterility and could
spread infection. Helman Stern of the BGU Department of Industrial
Engineering and Management says the Gestix gesture recognition system
functions in two stages: The initial stage is a calibration stage where
the machine recognizes the surgeons' hand gestures, and the second stage is
where surgeons must learn and implement eight navigator gestures, which
involves rapidly moving their hands in and out of a "neutral area." Gestix
users can zoom in and out by moving the hand clockwise or counterclockwise,
and to prevent unintentional signals from being read, users can enter a
"sleep" mode by dropping their hands. The gestures are read by a camera
positioned about a large flat-screen monitor. The system runs on an Intel
Pentium processor and a Matrox Standard II video-capturing device. Helman
and Yael Edan, another principal investigator, have used hand gesture
recognition as part of an interface to evaluate different aspects of
interface design on performance in a variety of telerobotic and
teleoperated systems, and ongoing research aims to expand this work to
include additional control modes, such as voice, to create a multimodal
telerobotic control system.
Click Here to View Full Article
to the top
Exciton-Based Circuits Eliminate a 'Speed Trap' Between
Computing and Communication Signals
University of California, San Diego (06/18/08)
University of California, San Diego physicists have demonstrated that
excitons, particles that emit a flash of light as they decay, could be used
for a new method of computing that is better suited to fast communication.
Integrated circuits currently use electrons to send signals needed for
computation, but almost all communications devices use light, or photons,
to send signals. The need to convert the signaling language from electrons
to photons limits the speed of electronic devices. UCSD physics professor
Leonid Butov and colleagues have built several exciton-based transistors
that could be used in a new type of computer. "Our transistors process
signals using excitons, which like electrons can be controlled with
electrical voltages but unlike electrons transform into photons at the
output of the circuit," says Butov. "This direct coupling of excitons to
photons bridges a gap between computing and communications." Excitons are
created by light in a semiconductor like gallium arsenide, which separates
a negatively charged electron from a positively charged "hole." If the
pair remains linked, it forms an exciton, and when the electron recombines
with the hole, the exciton decays and releases its energy as a flash of
light. The researchers used a special type of exciton where the electron
and its hole are confined to different "quantum wells" separated by several
nanometers, which creates an opportunity to control the flow of excitons
using voltage supplied electrodes. The voltage fates create an energy bump
that can halt the movement of excitons or allow them to flow, and removing
the energy barrier allows the exciton to travel to the transistor output
and transform to light, which could be fed directly into a communication
circuit, eliminating the need to convert the signal.
Click Here to View Full Article
to the top
Google Sponsors Scholarships for Grace Hopper Celebration
of Women in Computing Conference
Business Wire (06/18/08)
The Anita Borg Institute for Women and Technology recently announced that
Google will fund more than 50 sponsorships for women to attend the 8th
Grace Hopper Celebration of Women in Computing Conference, which offers
students, specifically female computer science students, a unique
opportunity to meet each other and network with other computer science
students and professionals with a passion for their work and for increasing
diversity in the field. The scholarships are available through a variety
of areas, including the Google Women of Color Scholarship; the Google
Global Community Scholarship, which is awarded to a group of international
computer science students studying at universities outside the United
States; Change Agent, which is awarded to accomplished technical women from
emerging countries; and Google Anita Borg Scholars, which will award
recipients with $10,000 in addition to registration and travel to the Grace
Hopper Celebration. Academic scholarships for computer science students
and travel scholarships to allow students to attend conferences like the
Celebration of Women in Computing serve to support the goal of retaining
and increasing the number of women and other underrepresented groups in the
field of computer science.
Click Here to View Full Article
to the top
Standardization of Rule Based Technologies Moving Forward
to Enable the Next Generation Web
Pressemitteilung Web Service (06/11/2008)
RuleML is following up its first industry-oriented gathering last year
with the International RuleML Symposium on Rule Interchange and
Applications (RuleML-2008). This year's event is scheduled for Oct. 30-31
at the Buena Vista Palace in the Walt Disney World Resort in Orlando, Fla.,
and will give business and technology professionals, researchers, and
standardization representatives another opportunity to focus on the
increasing performance and applicability of rule technologies. The
international umbrella organization for Web rule research, standardization,
and adoption will make every topic related to rules a focus of the
symposium, from engineering and use of rule-based systems, the integration
of rules and other Web technologies, languages and frameworks for rule
representation and processing, rule-related web standards, to the
interoperation of rule-based systems and the incorporation of rule
technology into enterprise architectures. RuleML-2008 will offer
peer-reviewed paper presentations, invited talks, software demonstrations,
and social events. There will be a challenge session that gives
participants the opportunity to show their commercial and open source
tools, use cases, and applications, and to win prizes for best
applications. The symposium will be co-located with the Business Rules
Forum to bring greater attention to the connection between rule and
business logic technologies. ACM is among the sponsors and partner
organizations that support RuleML-2008.
Click Here to View Full Article
to the top
Microsoft Researcher Tackles Forest Data to Fight Climate
Change
IT News Australia (06/13/08) Tay, Liz
The computational challenges of environmental research have become a focus
of Microsoft Computational Science Research, a new unit that has brought
ecologists, biologists, neuroscientists, mathematicians, and computer
scientists together to study climate change and other issues. "These
computational challenges are huge, and large software companies are one of
the only places where we're going to find the knowledge and resources to
address them," says ecologist Drew Purves. A research scientist at
Microsoft Research Cambridge, Purves believes new computer models are
needed to better predict environmental change. Purves, ecologist Stephen
Pacala at Princeton University, and research colleagues in Madrid, Spain,
are the authors of two research papers that recently appeared in the
international journal Science. They favor new algorithms that would
perform large numbers of calculations on ecological data sets, and would
account for biodiversity. "Of course there are lots of these algorithms
used every day in all kinds of fields [such as] actuaries or search
engines, but we have developed some ideas for new algorithms that we think
will be useful for ecology and, hopefully, in other areas," says Purves.
Click Here to View Full Article
to the top
Right on Cue
The Engineer (06/15/08) Vol. 293, No. 7749, P. 7; Baker, Berenice
Researchers at Portsmouth University are utilizing artificial intelligence
software developed by Neuron Systems to identify sounds that might indicate
a crime in progress and trigger CCTV cameras to swing towards the source.
The three-year, EPSRC-sponsored project aims to adapt the software, which
currently recognizes visual patterns, for sound cues. David Brown of
Portsmouth's Institute of Industrial Research says the system would employ
fuzzy logic to identify a type of noise, and notes that "we are looking for
templates of sound--riser response, shapes of sound." Among the project's
major challenges is keying the system for real-time responses to anomalous
sounds on a scale that is comparable to human response times, and the
software will work in tandem with CCTV-based human motion analysis that
Brown's institute developed as well. "If the camera is pointing in a
direction because an aggressive sound has been identified, the motion
software can identify whether a person is punching another or running away
from the scene," Brown remarks. The project could also help address the
challenge of sifting through hours of security video to identify a
particular action or object. By the project's conclusion, the team hopes
to have produced algorithms that can be used within a commercial software
suite, with each generation of algorithms growing in sophistication as the
project progresses.
Click Here to View Full Article
to the top
Michigan Tech Physicist Models Single Molecular
Switch
Michigan Technological University (06/16/08) Goodrich, Marcia
A team of Michigan Technological University researchers led by physicist
Ranjit Pati have developed a model to explain the mechanism behind the
single molecular switch, widely considered to be computing's Holy Grail.
If worked out experimentally, the model could help explode Moore's Law and
revolutionize computing technology. The fabled molecular switch would
essentially involve replacing the current generation of transistors with
molecules, which would allow 1 trillion switches to fit onto a
centimeter-square chip. In 1999, a team of researchers at Yale University
published a description of the first molecular switch, but scientists have
been unable to replicate their discover or explain how it worked; Pati
believes he and his team have discovered the mechanism behind the switch.
Applying quantum physics, Pati and his group developed a computer model of
an organometallic molecule firmly bound by two gold electrodes. After a
current was turned on, the current increased along with the voltage, until
it rose to a miniscule 142 microamps and suddenly, and counterintuitively,
dropped, a phenomenon known as negative differential resistance (DNR). Up
until the 142-microamp switch, the molecule's cloud of electrons had been
whizzing about the nucleus in equilibrium, similar to planets orbiting the
sun. However, that state fell apart under the higher voltage, and the
electrons were forced into a different equilibrium, a process known as
"quantum phase transition." Pati says he never thought this result could
happen. A molecule capable of exhibiting two different phases when
subjected to electric fields could function as a switch, with one phase
acting as a "zero" and the other as the "one" to form the foundation of
digital electronics. Pati and other scientists are now working to test the
model experimentally.
Click Here to View Full Article
to the top
Researchers Develop Ultra Low-Cost Plastic Memory
University of Groningen (06/16/08)
Zernike Institute of Advanced Materials researchers at the University of
Groningen in the Netherlands have developed a technology for a plastic
ferro-electric diode, similar to technology used in flash memory chips,
which the researchers believe will lead to a breakthrough in the
development of ultra low-cost plastic memory material. Like flash, plastic
memory can retain data without being connected to a power source. The
researchers expect the new technology to lead to the development of
products comparable with, and possibly even more significant than, flash
memory. In 2005, a joint team of researchers from the University of
Groningen and Philips successfully integrated a ferro-electric polymer into
a plastic transistor; the ferro-electric material can operate as a
non-volatile memory because the material can be switched between two
different stable states when exposed to a voltage pulse, while the
disadvantage of such a transistor is that three connections are needed for
programming and reading memory, making fabrication more complicated. The
challenge the researchers faced was to create a memory component with only
two connections, a diode. The breakthrough was based on a radical new
concept that, instead of stacking a layer of semiconducting material on a
layer of ferro-electric material, uses a mixture of the two materials. The
ferro-electric characteristic of the mixture is then used to direct current
through the semi-conducting part of the mixture. The new memory diode can
be programmed quickly, retains data for long periods, and operates at room
temperature. Additionally, the voltages needed for programming are low
enough for the diode to be used in commercial applications, and the
material can be manufactured at a low cost using large-scale industrial
production techniques.
Click Here to View Full Article
to the top
New Intrusion Tolerance Software Fortifies Server
Security
George Mason University (06/16/08)
Researchers at George Mason University are taking a different approach to
intrusion detection and prevention. Arun Sood, professor of computer
science and director of the Laboratory of Interdisciplinary Computer
Science, and Yin Huang, senior research scientist in the Center for Secure
Information Systems, accept the likelihood of someone trespassing on
computer servers, but believe limiting the time of continuous connection to
the Internet can serve as an additional layer of defense. Sood and Huang
have developed Self Cleansing Intrusion Tolerance (SCIT), and they use
virtualization technology to create duplicate servers. The idea is to
periodically cleanse an online server and restore it to a known clean
state, regardless of whether an intrusion has been detected. Regular
cleansings occur in sub-minute intervals. "SCIT interrupts the flow of
data regularly and automatically, and the data ex-filtration process is
interrupted every cleansing cycle," says Sood. "Thus, SCIT, in partnership
with intrusion detection systems, limits the volume of data that can be
stolen."
Click Here to View Full Article
to the top
Robot Asimo Can Understand Three Voices at Once
NewScientistTech (06/10/08) Barras, Colin
Researchers in Japan have developed new software that enables the advanced
humanoid robot Asimo to understand three voices at once. Hiroshi Okuno at
Kyoto University and Kazuhiro Nakadai at the Honda Research Institute in
Saitama were inspired by the legend of Prince Shotoku, who is said to have
had the ability to listen to the petitions of 10 people at the same time.
They note that the "Prince Shotoku Effect" is very different from the
"cocktail party effect," which involves focusing on a single voice while
one can hear other voices. The new software, HARK, makes use of an array
of eight microphones to determine where each voice is coming from and
isolate it from other sound sources, gauges the reliability of its
extracted individual voice, and uses speech-recognition software to decode
it. Asimo has used HARK to judge rock-paper-scissors contests, and has had
an accuracy rate of 70 percent to 80 percent. Okuno and Nakadai plan to
increase the number of voices and complexity of sentences that HARK can
understand. They presented their research at the 2008 IEEE International
Conference on Robotics and Automation in Pasadena, Calif., in May.
Click Here to View Full Article
to the top
As Chips Go Multicore, Researchers Strive to Parallelize
More Apps
SearchDataCenter.com (06/11/08) Botelho, Bridget
Stanford University and some of the largest organizations in the computing
industry recently announced the creation of the Pervasive Parallelism Lab
(PPL), intended to find a way for software developers to easily parallelize
applications for multicore processing. Sun Microsystems, Advanced Micro
Devices, NVIDIA, IBM, Hewlett-Packard, and Intel will support Stanford
computer scientists and electrical engineers as they conduct their research
and development efforts. The PPL has a budget of $6 million over the next
three years to research and develop a top-to-bottom parallel computing
system, from hardware to user-friendly programming languages, which will
allow developers to exploit parallelism automatically. As part of another
effort, Microsoft and Intel expect to invest a combined $20 million over
the next five years to fund similar research at the University of
California, Berkeley, and the University of Illinois at Urbana-Champaign.
Until recently, multicore processors were too expensive for any application
other than supercomputers, where multithreading is common. The limited use
of multithreading means few software programmers learned how to design
software that uses parallelism to exploit multiple cores, according to
Stanford's PPL research director Kunle Olukotun. "We are working with
application developers to provide solutions for their applications,"
Olukotun says. "After we figure out how to parallelize specific
applications, we will work on doing it in a more general context. My hope
is that our efforts will pave the way for programmers to create software
for applications such as artificial intelligence and robotics, business
data analysis, virtual worlds and gaming."
Click Here to View Full Article
to the top
Melding Mind and Machine
The Institute (06/06/08) Vol. 32, No. 2, P. 5; Riezenman, Michael J.
Giving paralysis victims mobility and a means to communicate is one of the
goals of research into brain-machine interfacing (BMI), in which an
artificial system senses and analyzes neural signals, and then translates
those signals into movement. Ideally the sensory apparatus should be
noninvasive, but IEEE members point out that such an approach can yield
poor signal-to-noise ratio. Conversely, while surgical implantation of the
interface may generate more accurate readings, the risk of infection and
tissue trauma is a clear concern. Efforts to refine both types of BMI are
showing progress, while still another avenue of research involves
experimentation with an electrocorticographic method that positions a small
electrode array on the cerebral cortex, yielding signals that suffer a lot
less attenuation than EEG signals while manifesting a higher
signal-to-noise ratio. Another significant BMI issue is keeping an
implantable device's power consumption to a minimum so that battery life is
maximized and the heating of the cerebral tissue is trivial. Minimizing
the bandwidth occupied by the data being transmitted from the implanted
device to the outside world is one possible approach, and a University of
Florida professor devised a bandwidth-saving scheme that samples faster
when the signal amplitude is large and slower when it is small.
Click Here to View Full Article
to the top
