GT Spectrum

Reports from the IT horizon

by / October 28, 2003
Sandia Develops Cognitive Machines
ALBUQUERQUE, N.M. -- A new "smart" machine that could fundamentally change how people interact with computers is being tested at the Department of Energy's Sandia National Laboratories.

For the past five years, a team led by Sandia cognitive psychologist Chris Forsythe has been developing cognitive machines that accurately infer user intent, remember experiences with users, and allow users to call on simulated experts to help them analyze situations and make decisions.

The initial goal of the work was to create a "synthetic human" -- a software program/computer that could think like a person.

"The benefits from this effort are expected to include augmenting human effectiveness and embedding these cognitive models into systems like robots and vehicles for better human-hardware interactions," said John Wagner, manager of Sandia's Computational Initiatives Department. "We expect to model, simulate and analyze humans and societies of humans for Department of Energy, military and national security applications."

Massive computers that could compute large amounts of data were available, said Forsythe. "But software that could realistically model how people think and make decisions was missing," he said.

There were two significant problems with previous modeling software. First, the software did not relate to how people actually make decisions -- it followed logical processes, which people don't necessarily do. People make decisions based, in part, on experiences and associative knowledge. Software models of human cognition also did not take into account factors such as emotions, stress and fatigue.

In an early project, Forsythe developed the framework for a computer program that used both factors.

Follow-up projects developed methodologies that allowed the knowledge of a specific expert to be captured in computer models and provided synthetic humans with episodic memory -- memory of experiences -- so they might apply their knowledge of specific experiences to solving problems in a manner that closely parallels what people do.

"Systems using this technology are tailored to a specific user, including the user's unique knowledge and understanding of the task," said Forsythe.

Work on cognitive machines started in 2002 with a contract from the Defense Advanced Research Projects Agency (DARPA) to develop a machine that can infer an operator's cognitive processes. This capability provides the potential for systems that augment the cognitive capacities of an operator through "discrepancy detection."

In discrepancy detection, the machine uses an operator's cognitive model to monitor its own state, detecting discrepancies between the machine's state and the operator's behavior.

Early this year, work began on Sandia's Next Generation Intelligent Systems Grand Challenge project.

"The goal of this Grand Challenge is to significantly improve the human capability to understand and solve national security problems, given the exponential growth of information and very complex environments," said Larry Ellis, the principal investigator.

"It's entirely possible," said Sandia's Forsythe, "that these cognitive machines could be incorporated into most computer systems produced within 10 years." -- Sandia National Laboratories

IBM Delivers World's Most Powerful Linux Supercomputer
TOKYO -- Japan's largest national research organization announced at the end of July that it ordered an IBM eServer Linux supercomputer that when completed, will deliver more than 11 trillion calculations per second, making it the world's most powerful Linux-based supercomputer.

It is expected to be more powerful than the Linux cluster at Lawrence Livermore National Laboratory in Livermore, Calif., which is currently ranked the third most powerful supercomputer in the world, according to the independent TOP500 List of Supercomputers.

The plan is to integrate the supercomputer with other non-Linux systems to form a massive, distributed computing grid -- enabling collaboration between corporations, academia and government -- to support various research including grid technologies, life sciences, bioinformatics and nanotechnology.

The system -- with a total of 2,636 processors -- will include 1,058 eServer 325 systems. The powerful new supercomputer will help Japan's National Institute of Advanced Industrial Science and Technology (AIST), known worldwide for its leading research in grid technologies, to accelerate research using grid technology for a wide variety of projects.

These projects include the search for new materials to be used for superconductors and fuel cell batteries, and the search for new compounds that could be the basis for a cure for various malignant diseases.

Each new IBM eServer 325 system delivered to AIST contains two powerful AMD Opteron processors in a 1.75" rack mounted form factor. AIST will run SuSE Linux Enterprise Server 8 on the supercomputer. The grid will incorporate the Globus Toolkit 3.0 and the Open Grid Services Infrastructure.

The grid is also planned to link heterogeneous and geographically dispersed computing resources, including servers, storage and data, allowing researchers to collaborate. The eServer 325 systems are designed to run either Linux or Windows operating systems, and 325 can run both 32-bit and 64-bit applications simultaneously. -- IBM

FDA Approves Stair-Climbing Wheelchair
WASHINGTON, D.C. -- The U.S. Food and Drug Administration approved a battery-powered wheelchair in August that relies on a computerized system of sensors, gyroscopes and electric motors, which allow indoor and outdoor use on stairs, and on level and uneven surfaces.

The FDA expedited review of the product -- the Independence iBOT 3000 Mobility System -- because it has the potential to benefit people with disabilities. An estimated 2 million people in the United States use wheelchairs.

FDA Commissioner Mark B. McClellan said, "It can help improve the quality of life of many people who use wheelchairs by enabling them to manage stairs, reach high shelves and hold eye-level conversations."

Four-wheel drive enables users to traverse rough terrain, travel over gravel or sand, go up slopes and climb 4-inch curbs. For use on stairs, two sets of drive wheels rotate up and over each other to climb up or down, one step at a time. Because of its unique balancing mechanism, the wheelchair remains stable and the seat stays level during all maneuvers.

The user can push a button to operate the wheelchair in several different ways.

To climb up stairs, the occupant backs up to the first step, holds onto the stair railing, shifts his weight over the rear wheels, which causes the chair to begin rotation of the front wheels over the rear wheels and then down to the first step. As the user shifts his weight backward and forward, the chair senses this and adjusts the wheel position to keep his center of gravity under the wheels. The chair ascends stairs backward and descends forward (the user always faces down the stairs).

To reach high shelves or hold eye-level conversations with people who are standing, the occupant shifts his weight over the back wheels, so the iBOT lifts one pair of wheels off the ground and balances on the remaining pair. The user then presses a button to lift the seat higher.

People must weigh no more than 250 pounds and must have the use of at least one arm to operate the chair. They also must have good judgment skills to discern which obstacles, slopes and stairs to prevent serious falls. Users must be capable of some exertion when climbing stairs in the wheelchair by themselves. However, for users who cannot tolerate such exertion, there is a feature that allows someone else to hold onto and tilt the chair's back to allow it to climb up or down stairs.

Physicians and other health professionals must undergo special training to prescribe the iBOT. The chair must be calibrated to the patient's weight; and patients must be trained in its use and pass physical, cognitive and perception tests to prove they can operate it safely.

The FDA approved the wheelchair based on a review of extensive bench testing of the product conducted by the manufacturer -- Independence Technology of Warren, N.J. -- and on a clinical study of its safety and effectiveness. Approval was also based on recommendation of the Orthopedic and Rehabilitation Devices Panel of the FDA's Medical Devices Advisory Committee.

The firm performed a wide range of tests on the chair, including mechanical, electrical, performance, environmental and software testing.

In the pivotal clinical study, 18 patients -- mostly people with spinal cord injury -- were trained to use the iBOT. They test-drove for two weeks to allow researchers to check maneuverability, falls and other problems compared to those encountered with their regular wheelchairs. They also tested it going up hills, over bumpy sidewalks, crossing curbs, reaching shelves and climbing stairs. Twelve patients could climb up and down stairs alone with the iBOT and the other six patients used an assistant. When these same 18 patients used their regular wheelchairs, one patient could "bump" down stairs, but no one could go up just one step.

During the pivotal study, three patients fell out of the iBOT and two fell out of their own wheelchairs. None of the falls occurred on stairs. Two patients experienced bruises while using the iBOT.

As a condition for approval, the manufacturer agreed to provide periodic reports to the FDA to document the chair's usage, functioning and any patient injuries. The manufacturer also said the iBOT will be available throughout the next few months in strategically located clinics across the country. -- The U.S. Food and Drug Administration
Shane Peterson Associate Editor