Articles tagged with Artificial Neural Networks
Researchers at UC Santa Barbara demonstrate a simple artificial neural circuit that performs image classification, using memristor technology to achieve brain-like efficiency. The breakthrough has potential applications in medical imaging, navigation systems, and search technologies.
A study published in PLOS Computational Biology found that an artificial deep neural network performs as well as the primate brain at object recognition. This achievement could pave the way for significant advancements in artificial intelligence and our understanding of primate visual processing.
Researchers developed a neuromorphic system that can carry out complex sensorimotor tasks in real time, exhibiting cognitive abilities. The system combines artificial neurons into networks that implemented neural processing modules, closely resembling mammalian brain structures.
Artificial neural networks (ANNs) improve patient survival prediction in advanced brain cancers, with a pooled voting method correctly predicting risk of death within one year in 84% of patients. A new proposal calls for medical professionals and specialty societies to play an increased role in evaluating 'off-label' uses of medications.
A team of researchers has developed an artificial neural network model to predict the location of fossil sites. The software uses satellite imagery and maps to identify productive areas in the Great Divide Basin, Wyoming, and has already accurately pinpointed 79% of known fossil sites.
Researchers at Caltech created an artificial neural network out of DNA, exhibiting brain-like behavior by recalling memories based on incomplete patterns. The DNA-based neural network consists of four artificial neurons made from 112 distinct DNA strands and demonstrated correct responses in a mind-reading game.
A team of researchers has developed a new model that uses artificial neural networks to predict adverse drug reactions (ADRs) among 10,000 observations with 99.87% accuracy. The technology has the potential to save lives by identifying potential ADRs at an early stage of drug development and marketing.
A team of chemists used artificial neural networks to analyze tea leaves' mineral content and identify the type of tea. The technique achieved a high accuracy rate, allowing for clear differentiation between white, green, black, Oolong, and red tea varieties.
NeuFlow is a new supercomputer that processes tens of megapixel images in real time, allowing for rapid object recognition. The system has the potential to enable self-driving cars by recognizing various objects on the road, such as other cars, people, and stoplights.
Researchers developed an artificial neural network (ANN) to evaluate symptoms and predict endocarditis diagnoses. The ANN achieved high accuracy in distinguishing between infected and non-infected cases, eliminating the need for invasive transesophageal echocardiography.
Researchers developed a neural network system to classify music genres, such as cha-cha-cha, jive, and tango, with varying degrees of success. The approach combines the strengths of two existing methods and uses a neural network to analyze beat and tempo, outperforming other classification techniques.
Researchers have developed an analytical method using Empirical Mode Decomposition (EMD) to classify a wider range of heart sounds than skilled physicians can. The EMD system, trained with AI algorithms, outperforms conventional methods in identifying murmurs and other anomalies.
A study published in the World Journal of Gastroenterology found that artificial neural networks can accurately predict thyroid disease in patients with atrophic body gastritis. The analysis of 253 ABG patients revealed a high accuracy rate of 76%, correctly identifying 82% of patients with thyroid disease.
A new theory proposes that learning skills, such as flying, accelerates the evolution of innate abilities in birds by creating a latent memory that reduces the need for future generations to learn. This is achieved through the use of distributed representations in neural networks, which allows for faster evolution of adaptive behaviors.
Jeff Elman's work in connectionism and artificial neural networks has led to breakthroughs in speech perception, language processing, and cognition. His creation of the TRACE model and Simple Recurrent Network has been widely used to simulate human behavior.
Ames Laboratory researchers have created an AI-powered system that can detect secret files hidden in digital images using steganalysis. The system, utilizing artificial neural networks (ANNs), has been trained on a database of over 10,000 images and achieved high accuracy rates.
A new online tool provides personalized 10-year survival predictions for prostate cancer patients, taking into account age, race, co-morbidities, and treatment type. The model also highlights the significant impact of co-morbidities on long-term survival rates.
A prototype system using NIST-patented microheater technology and artificial neural networks can reliably identify trace amounts of toxic gases. The sensors can detect compounds like sulfur-mustard gas and nerve agents at levels below 1 part per million.
Researchers have developed an AI system that can recognize various marine species, including fish and potential threats. The system, called Fetch2, has successfully identified two species - jacks and sharks - using side scan sonar data and neural networks, paving the way for autonomous surveillance of coastlines and harbors.
Researchers at Pacific Northwest National Laboratory are developing TEDANN to predict failures and abnormal operations in M1 Abrams main battle tanks' turbine engines. The technology uses diagnostic engineering, artificial neural networks, and model-based decision algorithms to enhance tank readiness while reducing costly engine failures.