DUAL takes AI to the next level
A brain-inspired computing architecture speeds up complex data processing by running its algorithms inside its memory, significantly saving time and energy.
2021-01-02
(Press-News.org) "Today's computer applications generate a large amount of data that needs to be processed by machine learning algorithms," says Yeseong Kim of Daegu Gyeongbuk Institute of Science and Technology (DGIST), who led the effort.
Powerful 'unsupervised' machine learning involves training an algorithm to recognize patterns in large datasets without providing labelled examples for comparison. One popular approach is a clustering algorithm, which groups similar data into different classes. These algorithms are used for a wide variety of data analyses, such as identifying fake news on social media, filtering spam in our e-mails, and detecting criminal or fraudulent activity online.
"But running clustering algorithms on traditional cores results in high energy consumption and slow processing, because a large amount of data needs to be moved from the computer's memory to its processing unit, where the machine learning tasks are conducted," explains Kim.
Scientists have been looking into processing in-memory (PIM) approaches. But most PIM architectures are analog-based and require analog-to-digital and digital-to-analog converters, which take up a huge amount of the computer chip power and area. They also work better with supervised machine learning, which includes labelled datasets to help train the algorithm.
To overcome these issues, Kim and his colleagues developed DUAL, which stands for digital-based unsupervised learning acceleration. DUAL enables computations on digital data stored inside a computer memory. It works by mapping all the data points into high-dimensional space; imagine data points stored in many locations within the human brain.
The scientists found DUAL efficiently speeds up many different clustering algorithms, using a wide range of large-scale datasets, and significantly improves energy efficiency compared to a state-of-the-art graphics processing unit. The researchers believe this is the first digital-based PIM architecture that can accelerate unsupervised machine learning.
"The existing approach of the state-of-the-arts in-memory computing research focuses on accelerating supervised learning algorithms through artificial neural networks, which increases chip design costs and may not guarantee sufficient learning quality," says Kim. "We showed that combining hyper-dimensional and in-memory computing can significantly improve efficiency while providing sufficient accuracy."
INFORMATION:
[Attachments] See images for this press release:
ELSE PRESS RELEASES FROM THIS DATE:
2021-01-02
DALLAS - Dec. 30, 2020 - When fat cells in the body are stuffed with excess fat, the surrounding tissue becomes inflamed. That chronic, low-level inflammation is one of the driving factors behind many of the diseases associated with obesity. Now, UT Southwestern scientists have discovered a type of cell responsible, at least in mice, for triggering this inflammation in fat tissue. Their findings, published in Nature Metabolism, could eventually lead to new ways to treat obesity.
"The inflammation of fat cells in obese individuals is linked to many of the ...
2021-01-02
Lithium is an energy-critical element that is considered to be a geopolitically significant resource. However, the supply of lithium may not be enough to meet continuously increasing demand. As a result, scientists are looking for new ways to extract lithium ions.
Ion selective membranes have already been used extensively for water treatment and ion sieving in electrodialysis technology. However, conventional membranes exhibit low and useless Li+ selectivity, making them insufficient for meeting industry requirements.
Chinese scientists have recently made progress in the preparation and application of a bioinspired ...
2021-01-02
Arctic sea ice has been rapidly declining in recent decades, and changes in arctic sea ice can have a significant impact on global weather and climate through interactions with the atmosphere and oceans. In addition, the Arctic shipping routes are a shortcut to connect the major countries in the Northern Hemisphere. The Arctic region is also rich in natural resources and biological resources. Simulation of the Arctic sea ice could provide valuable information for Arctic shipping as well as climate studies, and it is therefore urgent to evaluate the ability to simulate Arctic sea ice and diagnose the sources of simulation errors.
To address the issue of error source identification, ...
2021-01-02
Experts in optical physics have developed a new way to see inside living cells in greater detail using existing microscopy technology and without needing to add stains or fluorescent dyes.
Since individual cells are almost translucent, microscope cameras must detect extremely subtle differences in the light passing through parts of the cell. Those differences are known as the phase of the light. Camera image sensors are limited by what amount of light phase difference they can detect, referred to as dynamic range.
"To see greater detail using the same image sensor, we must expand the dynamic range so that we can ...
2021-01-02
Activating an immune signaling pathway best known for fighting viral and bacterial infections can boost the ability of genetically engineered T cells to eradicate breast cancer in mice, according to a new study by researchers at the University of North Carolina. The study, to be published December 31 in the Journal of Experimental Medicine (JEM), suggests that CAR T cells, which are already used to treat certain blood cancers in humans, may also be successful against solid tumors if combined with other immunotherapeutic approaches.
Chimeric antigen receptor (CAR) T cells are a type of white blood cell that have been genetically engineered to recognize ...
2021-01-02
The Asian tiger mosquito does not pose a major risk for Zika virus epidemics, according to a study published December 31 in the open-access journal PLOS Pathogens by Albin Fontaine of the Institut de Recherche Biomédicale des Armées, and colleagues.
Zika virus has triggered large outbreaks in human populations, in some cases causing congenital deformities, fetal loss, or neurological problems in adults. While the yellow fever mosquito Aedes aegypti is considered the primary vector of Zika ...
2021-01-02
A novel computational drug screening strategy combined with lab experiments suggest that pralatrexate, a chemotherapy medication originally developed to treat lymphoma, could potentially be repurposed to treat Covid-19. Haiping Zhang of the Shenzhen Institutes of Advanced Technology in Shenzhen, China, and colleagues present these findings in the open-access journal PLOS Computational Biology.
With the Covid-19 pandemic causing illness and death worldwide, better treatments are urgently needed. One shortcut could be to repurpose existing drugs that were originally developed to ...
2021-01-02
Multiple bouts of blood feeding by mosquitoes shorten the incubation period for malaria parasites and increase malaria transmission potential, according to a study published December 31 in the open-access journal PLOS Pathogens by Lauren Childs of Virginia Tech, Flaminia Catteruccia of the Harvard T.H. Chan School of Public Health, and colleagues. Given that mosquitoes feed on blood multiple times in natural settings, the results suggest that malaria elimination may be substantially more challenging than suggested by previous experiments, which typically involve a single blood meal.
Malaria ...
2021-01-02
The discovery of new drugs is vital to achieving the eradication of neglected tropical diseases (NTDs) in Africa and around the world. Now, researchers reporting in PLOS Neglected Tropical Diseases have identified traditional Ghanaian medicines which work in the lab against schistosomiasis, onchocerciasis and lymphatic filariasis, three diseases endemic to Ghana.
The major intervention for NTDs in Ghana is currently mass drug administration of a few repeatedly recycled drugs, which can lead to reduced efficacy and the emergence of drug resistance. Chronic infections of schistosomiasis, onchocerciasis and lymphatic filariasis ...
2021-01-02
In the ongoing arms race between humans and the parasite that causes malaria, Taane Clark and colleagues at the London School of Hygiene and Tropical Medicine (LSHTM) report that new mutations that enhance resistance to a drug used to prevent malaria in pregnant women and children are already common in countries fighting the disease. The new results are published December 31 in PLOS Genetics.
Malaria causes about 435,000 deaths each year, primarily in young children in sub-Saharan Africa. Despite a long-term global response, efforts to control the disease are hampered by the rise of drug-resistant strains of the parasite species that cause malaria. Sulfadoxine-pyrimethamine (SP), ...
LAST 30 PRESS RELEASES:
[Press-News.org] DUAL takes AI to the next level
A brain-inspired computing architecture speeds up complex data processing by running its algorithms inside its memory, significantly saving time and energy.
