NSF new program helps large laboratory transition ideas to the practical -
The "valley of death" is well known entrepreneurs - the lull between public funding research and industry support to prototypes and products. To address this problem, in 2013 the National Science Foundation has created a new program called InTrans to extend the life of the highest incidence NSF-funded research and help large laboratory transition ideas into practice.
Today, partnership with Intel Corporation, NSF announced the first InTrans price of $ 3 million to a team of researchers who design, specific IT technologies customizable field for use in health care.
work could lead to less exposure to harmful radiation during x-rays by accelerating the calculation part of medicine. It could also lead to specific treatments against the cancer patient.
Led by the University of California, Los Angeles, the research team includes experts in computer science and engineering, electrical engineering and medicine from Rice University and Oregon Health and Science University. The team comes mainly from IT Centre Domain-Specific (CDSC), which was supported by an NSF Expeditions in Computing Award in 09.
Expeditions, consisting of 5 years, $ 10 million price represent some of the most important investments being made by the NSF computer, information science and engineering (CISE) direction.
InTrans the grant today continues the research efforts funded by the shipping program to bring new technology to the point where it can be produced in a microchip manufacturing plant (or fob) for the mass market.
"We see the InTrans program as an innovative approach to public-private partnership and a means to improve the sustainability of research," said Farnam Jahanian, head of the Directorate of the NSF CISE. "We are delighted that Intel and NSF can partner to continue to support the development of specific equipment to the field and the transition of this excellent research fundament in real applications. "
in the project, the researchers looked au beyond parallelization (the process of working on a problem with more than one processor at the same time) and instead focused on specific own domain, a disruptive technology with the potential for improvement of orders of magnitude to important application specific computer systems to a domain to work effectively on specific issues. - in this case, medical imaging and DNA sequencing of tumors -. or set of problems with similar characteristics, which reduces solution time and lower costs
"We tried to create efficient computers energy that are more like the brain," explained Jason Cong, director of CDSC, a professor chancellor of computer and electrical engineering at UCLA, and the project manager.
"We do not really have a central unit for centralized processing there. If you look at the brain you have a region responsible for speech, another region for engine control, another area for vision. These are specialized "accelerators". We want to develop a system of this kind of architecture, where each accelerator can deliver hundred to a thousand times more effective than standard processors. "
The team plans to identify classes of applications that share the nuclei similar calculation, creating material that solves a common set of related problems with high efficiency and flexibility. This differs from dedicated circuits that are designed to solve one problem (such as those used in cell phones) or general purpose processors designed to solve all problems.
"the group asked another way to present the problem of computer specific area, which is :? How to determine the common features and support them effectively, "said Sankar Basu, program officer in the NSF." They developed a framework for the design of material for a domain they believe can be applied in many other areas too . "
the group selected medical imaging and specific cancer treatments patients - two important problems in health care - as test applications on which to create their design due to the impact of the health on the economy and the quality of national life.
medical imaging is now used to diagnose a variety of medical problems. However, diagnostic methods such as x-ray CT (computed tomography) scanners can exposing the body to a cumulative radiation, which increases the risk for long-term patient.
scientists have developed new algorithms for medical imaging which lead to less radiation exposure, but these have been forced because of a lack of computing power.
Due to their heterogeneous platform customizable, Cong and his team have done a major image reconstruction algorithms CT hundred times faster, which reduce the exposure of a subject to a radiation significantly. They presented their results in May 2014 at the International Symposium IEEE Custom field-programmable computing machines.
"The low dose scanner allows you to get a resolution similar to the standard CT, but the patient can get several times less radiation," said Alex Bui, a professor in the UCLA radiological sciences department and co-leader of the project. "Anything we can do to reduce that exposure will have a significant impact on health."
In theory, the technology also exists to determine the specific strain of cancer a patient has by DNA sequencing and use this information to design a specific treatment to the patient. However, it currently takes so long to sequence DNA that once the strain of a tumor is determined, the cancer has already mutated. With the material specific to the field, Cong believes rapid diagnosis and targeted treatments are possible.
"Power- and high cost-performance computing in these areas will have a significant impact on health in preventive medicine, therapeutic procedures and diagnostic procedures," said Cong.
" cancer genomics, in particular, has been hampered by the lack of open approaches, scalable and effective for aligning and quickly and accurately interpret the genome sequence data, "said Paul Spellman, a professor at OHSU, working on the personalized cancer treatment and served as another co-project manager.
"the ability to use hardware approaches to dramatically improve these speeds will facilitate rapid reversals in huge data sets will be required to deliver the accuracy of medicine. "
on the road, the team will work with Spellman and others at OHSU doctors to test the application of the equipment in a real environment.
" Intel excels in creating flat customizable iT Platforms optimized for HPC data, "said Michael C. Mayberry, vice president of technology from Intel and Manufacturing Group and Chairman of the corporate research Council. "These researchers are some of the leaders in the field of specific computer to the domain.
" This new effort allows us to maximize the benefits of Intel architecture. For example, we can ensure that the features of Intel Xeon processor are optimized under various accelerators for a specific field of application and across all architectural layers, "Mayberry said." Life Sciences and research health care will undoubtedly benefit from the performance, flexibility, energy efficiency and accessibility of this application. "
the program InTrans not only advances the important basic research and integrates into the industry, it also benefits society by improving technologies for medical imaging and cancer treatments, helping to prolong life.
"all the research to get to the stage where they are ready to make a direct impact on industry and society, but in our case, we're close," said Cong . "We are grateful for the support of NSF and are delighted to continue our research in the context of this model unique public-private financing. "
EmoticonEmoticon