Articles tagged with Neural Networks
Neuroscientists have uncovered a brain circuit that enables mice to rapidly escape to shelter when faced with a threat. The retrosplenial cortex and superior colliculus form a circuit that encodes the direction to a shelter, allowing mice to accurately orient and escape.
Researchers developed an AI-based neural network to detect early knee osteoarthritis from x-ray images, matching doctors' diagnoses in 87% of cases. This method could help reduce unnecessary examinations, treatments, and even knee joint replacement surgery.
A USTC team has made significant advancements in atomistic neural network (AtNN) representations for chemical dynamics simulations. By decomposing system properties into atomic contributions, AtNN can satisfy different symmetries and periodicities, achieving accurate and efficient molecular dynamics simulations of complex systems.
A review of over 100 scientific articles suggests that the safety of antidepressants during pregnancy is endorsed by science, but their effects on fetal neurodevelopment are poorly understood. Brazilian researchers propose using lab-grown mini-brains to investigate this impact and identify potential alterations.
Researchers at Incheon National University have developed an IoT-enabled, real-time object detection system for autonomous vehicles. The YOLOv3-based model achieved high accuracy (>96%) in detecting 2D and 3D objects, outperforming other state-of-the-art detection models.
Researchers at Baylor College of Medicine discovered that oxytocin drives the development of neural connections in adult-born neurons. Oxytocin triggers a signaling pathway that promotes synapse maturation, enabling these new neurons to function properly.
Researchers have developed a text-to-audio model that can generate coherent and relevant music and sound from text, opening up new avenues for creative application and exploration. The model employed data compression methods to improve output quality and reduce training time.
Research reveals that brain function networks are affected differently by aging, gender, and blood immune factors. The study found correlations between cytokine clock, brain shrinkage, and gender, with females having a faster ticking cytokine clock.
A new study reveals ketamine dramatically changes neuronal activity patterns in the cerebral cortex, turning off active neurons and turning on silent ones. This switch in brain activity may impact our understanding of ketamine's antidepressant effects and future research in neuropsychiatry.
Researchers at Bar-Ilan University have created functional, multi-layered neural networks that mimic elements found in the brain of mammals. The 'mini-brains' were generated through magnetic manipulations of neural progenitor cells in a three-dimensional collagen substrate.
Researchers at MIT have developed a new architecture for optical neural networks, which can perform complex linear algebra operations using light signals. The new design eliminates uncorrectable errors that limited the scalability of earlier systems, enabling larger networks with improved accuracy.
The researchers have developed an AI algorithm called M3GNet that can predict the structure and dynamic properties of any material. The algorithm was used to create a database of over 31 million yet-to-be-synthesized materials with predicted properties, facilitating the discovery of new technological materials.
Engineers at Tokyo Tech demonstrated a simple approach to improve AI classifier training using limited sensor data, increasing quality without extra cost. The proposed method promises to address the challenge of classification accuracy in real-world applications, where reliable answers are crucial.
Chung-Ang University researchers propose a new algorithm, MR-UCB and MR-APE, to tackle stochastic multi-armed bandit problems with heavy-tailed noise distributions. The methods guarantee minimal loss for worst-case scenarios with minimal prior information.
A Rutgers-led study found that a gene mutation associated with autism causes an overstimulation of brain cells, disrupting the normal information flow. The researchers used human stem cells and transplanting them into mouse brains to understand how the mutation affects brain development.
Children's brains exhibit rapid GABA boosts during learning sessions, allowing for efficient consolidation of new information. Adults' brains, on the other hand, struggle to stabilize learned knowledge due to reduced GABA levels, leading to slower learning rates.
A novel AI-based malware detection and classification system has been developed for 5G-enabled Industrial Internet of Things (IIoT) systems. The system achieved an accuracy rate of 97% on benchmark datasets, enabling the secure connection of applications such as smart cities and autonomous vehicles.
Researchers at CABBI used unmanned aerial vehicles with machine learning methods to select the best candidate genotypes in miscanthus breeding programs. The new method leverages high-resolution aerial imagery and three-dimensional neural networks to estimate crop traits such as flowering time, height, and biomass production.
Researchers at MIT have developed a machine-learning model that captures how sounds propagate through spaces, allowing for accurate visual renderings of rooms. This technique has potential applications in virtual and augmented reality, as well as improving AI agents' understanding of their environment.
Researchers developed an algorithm to extract bouton-like structures from calcium imaging data, identifying synchronized synapses during fictive locomotion. PQ-clustering outperformed other algorithms in mimicking synaptic activity patterns.
The US Department of Energy's Oak Ridge National Laboratory has developed a massive geographic dataset, USA Structures, using deep learning to forecast potential damage and accelerate emergency response. The dataset provides critical information on building outlines and attributes, enabling FEMA to prioritize response efforts.
A new project aims to create more race-inclusive AI in medicine by developing a distributed, inclusive data collection and learning framework that relies on smartphone apps. The framework uses federated learning, which allows models to be trained on device data while protecting user privacy.
Researchers at the University of Copenhagen have made a breakthrough in understanding schizophrenia by analyzing individual brain cells. The study identified specific neurons and networks affected by the disease, suggesting that targeting these areas could lead to new treatment options.
Researchers used a neural network model to study how brain regions interact during sleep. They found that the hippocampus and neocortex work together to convert fleeting information into long-term memory. The findings suggest that alternating between REM and slow-wave sleep stages is crucial for strong memory formation.
Researchers from University of Warsaw create spiking neuron using photons to mimic biological brain's behavior. This achievement paves the way for photonic neural networks that process information faster and more efficiently than conventional systems.
Researchers at MIT have developed a new method that uses optics to accelerate machine-learning computations on low-power devices. By encoding model components onto light waves, data can be transmitted rapidly and computations performed quickly, leading to over a hundredfold improvement in energy efficiency.
Researchers created a 3D electrode array that maps the locations and activity of up to 1 million potential synaptic links in living brains. The system uses recordings of millisecond-scale evolution of electrical pulses in tens of thousands of neurons, allowing for dense and accurate mapping of brain circuits.
A new multimodality prediction method jointly learned vision and sequence information for accurate solar wind speed prediction. The model achieved best performance in many metrics, including RMSE, MAE, and CORR, and demonstrated the effectiveness of each module in improving prediction accuracy.
A new CNN framework, PE-Net, is proposed for predicting machine remaining useful life (RUL) accurately. The framework uses a novel architecture with small-sized one-dimensional convolution kernels and deep networks to learn features from input time series signals.
Scientists have discovered a hidden structure in the connections between neurons in the brain, which is crucial for the stability of the neuronal network. By combining mathematical models with experimental recordings, researchers found that the relative ratios of connection strengths are more important than absolute values.
Scientists discovered a novel neural mechanism that accompanies unconsciousness, masking sensory inputs with spontaneous activity. The auditory cortex's response to sounds is indistinguishable from its own internal activity under anesthesia.
Researchers used machine learning algorithms to optimize climate models, increasing their accuracy and detail. By applying Generative Adversarial Networks (GANs) to climate simulations, the team was able to improve the models' ability to represent extreme precipitation events.
Physicists at the University of Basel have developed a computational shortcut for neural networks, allowing for faster calculation of optimal solutions without training. This breakthrough provides insight into neural network functioning and could help detect unknown phase transitions in materials and quantum systems.
Researchers from Xi'an Jiaotong-Liverpool University found that brain stimulation combined with a nose spray containing nanoparticles can improve recovery after ischemic stroke. The treatment increased cognitive and motor functions, and weighed more quickly than those treated with TMS alone.
Physicists used machine learning to compress a complex quantum problem into four equations, capturing the physics of electrons on a lattice with high accuracy. The approach could revolutionize how scientists investigate systems containing many interacting electrons and potentially aid in designing materials with sought-after properties.
Researchers have developed a simplified and fast optoretinography approach to measure retinal function, potentially accelerating the development of new treatments for eye diseases. The technique can collect data from three healthy subjects in just ten minutes and has been demonstrated to be reproducible.
Researchers at UTHealth Houston will create a coordinating unit for biostatistics, informatics, and engagement to advance knowledge about human brain neurons. The project aims to produce an open-access digital brain cell reference atlas to improve understanding of neurological functions and disorders.
Neuroscientists at Sainsbury Wellcome Centre discovered that individual neurons in the visual cortex of mice are modulated separately by attention and running. The study found that spatial attention and running influence neurons independently, with different dynamics.
Researchers have developed a new end-to-end neural network called Fourier Imager Network (FIN) that can speed up the reconstruction of holographic images. FIN works well on new types of samples not seen by the network during training, delivering high-quality images and improved computational speed.
Researchers at Linköping University discovered a biological mechanism that increases the strength of fear memories stored in the brain. This finding provides new knowledge on the mechanisms behind anxiety-related disorders and identifies shared mechanisms with alcohol dependence.
A new neural network model developed by Aalto University researchers can accurately predict the occurrence of fires in peatlands. The model identified a suite of interventions that would reduce fire incidence by 50-76%.
Scientists at UC San Diego have illuminated the role of key neurons that alter function in response to seasonal changes in light exposure. The study found that neurons change expression of neurotransmitters in response to day length stimuli, triggering behavioral changes.
A team from Ruhr-Universität Bochum developed a novel neural network that can classify tissue samples as containing tumors or not. The AI also generates an activation map showing where the tumor is detected, based on falsifiable hypotheses.
Researchers developed a new machine-learning method to understand force chains in jammed granular solids. The graph neural network approach can predict the position of force chains with high accuracy, even for complex systems and varying conditions.
Researchers discovered an inhibitory neuronal network in the brainstem that generates a synchronous rhythm, retracting mouse whiskers from their protracted positions. The oscillator consists of parvalbumin-expressing vIRt neurons firing bursts only during whisker retraction.
Seismologists have identified hundreds of thousands of microearthquakes along previously unknown fault structures in Oklahoma and Kansas, allowing them to map and measure earthquake clusters. The study found that nearly a 5% chance that a cluster would host a magnitude 4 or larger earthquake within a year if it reached a certain length...
Researchers at MIT developed an AI model that can detect Parkinson's disease from breathing patterns, using a neural network to assess the presence and severity of the condition. The device is non-invasive and can be used in patients' homes without any bodily contact.
Researchers built a two-stage warning system predicting solar flares within 48 hours via k-means clustering and neural networks. The model improved recall while increasing precision, but lost some positive sample information, affecting prediction accuracy.
A team of researchers developed a deep learning pipeline to analyze vascular system images of plants with high accuracy. The pipeline can detect vascular bundles, identify specific zones, and perform statistical analysis of traits in different stem internodes. This study has the potential to improve crop resilience and food security.
A proof-of-concept study developed three machine learning models to predict posttreatment recurrence in early-stage hepatocellular carcinoma patients. The models achieved high accuracy using imaging data alone, while combining clinical data did not significantly improve performance.
Despite DeepMind's neural network claiming superiority, scientists question its performance on predicting electron interactions in chemical systems. The BBB test set shows limited understanding of fractional-electron systems, raising concerns about the AI's ability to generalize.
Researchers developed a Flashover Prediction Neural Network (FlashNet) model to forecast deadly fire events, beating other AI-based tools with up to 92.1% accuracy across various building floorplans. The model's performance improved when given real-world data, highlighting its potential for saving firefighter lives.
Researchers at University of the Basque Country have developed a convolutional neural network to predict flow characteristics around flow control devices on wind turbines. The model achieves accurate results with minimal computational time, reducing errors compared to traditional CFD simulations.
Researchers at Princeton University used artificial intelligence to simulate ice formation by individual atoms and molecules with quantum accuracy. This breakthrough enables tracking of hundreds of thousands of atoms over longer timespans than previous simulations.
Researchers at the University of Tokyo have made a groundbreaking discovery about the development of the visual system in mice. By studying the neural networks in cortical and thalamic regions, they found that parallel pathways from the retinas form earlier than connections within cortical areas, challenging current understanding of co...
Physicists have created a way to simulate quantum entanglement between interacting particles using neural networks and fictitious 'ghost' electrons. This approach enables accurate predictions of molecule behavior, which could lead to breakthroughs in pharmaceutical development and material design.
Researchers at MIT have developed a machine-learning system that uses computer vision to monitor the 3D printing process and correct errors in real-time. The system successfully printed objects more accurately than other 3D printing controllers, enabling engineers to incorporate novel materials into their prints with ease.
Researchers identified regions in the cerebral cortex and thalamus with high bidirectional connections, which are thought to be essential for consciousness. The findings support the idea that these networks are key to pinpointing the location of consciousness.
The new AI system uses associative learning to detect similarities in datasets, reducing processing time and computational cost. By leveraging optical parallel processing and light signals, the system can identify patterns and associations more efficiently than conventional machine learning algorithms.
Researchers propose a novel paradigm using nanoscale nonlinear fluid dynamics to support recurrent neural networks in neuromorphic computing. The liquid film functions as an optical memory, enabling 'reservoir computing' capable of performing digital and analog tasks.