Articles tagged with Network Modeling
Researchers have developed a new AI model, AsymMirai, to predict 5-year breast cancer risk from mammograms. The model performs almost as well as the state-of-the-art Mirai, but with an interpretable reasoning process, making it a valuable adjunct to human radiologists.
This study leverages attention mechanisms based deep learning models to improve spike detection in greenhouse cultivated grain crops. The Swin Transformer model demonstrates superior accuracy, while the FRCNN-A provides a faster training alternative.
Researchers used Generative Adversarial Networks (GANs) to create synthetic weed images with high accuracy and realism. The CA-GAN model demonstrated superior performance in generating detailed plant features, such as leaf textures and shapes, while maintaining distinctiveness of each weed species.
Researchers developed chronological age prediction models by analyzing gene expression changes in the prefrontal cortex, identifying genes associated with aging and potential mechanisms. The models showed high correlation with age and demonstrated female and male-specific differences.
A new traffic signal concept, known as the 'white phase,' uses autonomous vehicles to expedite traffic flow at intersections. The concept has been shown to improve travel time for both pedestrians and vehicles, especially when autonomous vehicles make up a higher percentage of traffic.
A Kyoto University research group developed RENGE, a computational model to estimate gene regulatory networks in multicellular organisms. The method measures time-series gene expression and uses the proprietary model to infer regulatory dynamics.
A novel path choice model elucidates the influence of different attributes on pedestrians' local and global path choices. The study reveals that pedestrians locally perceive and react to environmental attributes like green views, guiding urban planning decisions.
A recent study found that large language models can accurately translate discharge summaries into more readable formats, improving patient understanding. However, initial implementation will require physician review due to safety concerns.
DomAda-FruitDet is a domain-adaptive anchor-free fruit detection model that achieves impressive average precision scores of up to 94.0% across various fruit datasets. The model effectively bridges the foreground and background domain gaps, enabling accurate and efficient auto-labeling in smart orchards.
The new AI model uses a visual map to explain each diagnosis, helping doctors follow its line of reasoning and check for accuracy. The tool aims to catch diseases in their earliest stages, making it easier on doctors and patients alike.
A recent Mayo Clinic study found that daylight saving time has a minimal impact on heart health. Researchers analyzed data from 36 million adults and found a slight increase in cardiovascular events during the spring and fall transitions, but this was deemed clinically insignificant.
A new AI tool, DeepGO-SE, successfully predicts the molecular functions of unknown proteins with high accuracy. This breakthrough enables researchers to analyze uncharacterized proteins, facilitating tasks such as drug discovery, metabolic pathway analysis, and disease associations.
A new study from MIT researchers found that doctors are less accurate in diagnosing skin diseases based on images of patients with darker skin. The researchers also discovered that an artificial intelligence algorithm can assist doctors in improving their diagnosis, although the improvements were more pronounced for lighter skin tones.
The team proposed a novel machine learning model with data augmentation, which accurately predicts the plastic anisotropic properties of wrought Mg alloys. The model showed significantly better robustness and generalizability than other models, paving the way for improved design and manufacturing of metal products.
A Brazilian study developed a method using AI to identify crop-livestock integration areas from satellite images. The approach can benefit agriculture by optimizing land use, diversifying farming activities, and encouraging sustainable practices.
Researchers have developed a novel optical neural network architecture that achieves nonlinear optical computation by precisely controlling ultrashort pulse propagation in multimode fibers. This approach streamlines the need for energy-intensive digital processes, achieving comparable accuracy with significantly reduced parameters.
A research group from Tohoku University Graduate School of Engineering has replicated human-like variable speed walking using a musculoskeletal model steered by a reflex control method reflective of the human nervous system. The breakthrough in biomechanics and robotics sets a new benchmark in understanding human movement.
Researchers from Purdue University used deep learning to generate growth models for various tree species, both with and without leaves. The AI models can produce complex tree models with detailed geometry, improving digital forest endeavors.
A new study using generative AI models simulated how the brain learns and remembers events, revealing how memories are re-constructed in our minds. The model showed how the hippocampus and neocortex work together to create efficient 'conceptual' representations of scenes, enabling us to both recall past experiences and imagine new ones.
A novel AI approach estimates nutritional value of harvested seed mixes, aiding farmers in crop yields and sustainable cultivation. The 'ESTI'METEIL' web component enables users to estimate seed composition and nutritional value from images.
The study introduces an innovative approach using an autoencoder network to predict fruit shapes from molecular data. GenoDrawing achieves more accurate predictions by targeting specific SNPs, outperforming randomly selected SNPs.
Researchers propose a simple model that accurately describes neuronal connectivity in various organisms, suggesting that general networking principles govern brain organization. The model also provides an unexpected explanation for clustering phenomenon in social interactions and can be extended to other types of networks.
A new study led by Dr. Richard Naud of the University of Ottawa's Faculty of Medicine tackles the mystery of neuronal response variability, controlling output with dendrites' inputs to the core and little antennas
Next-generation mobile networks are being optimized for increased data loads and faster speeds using AI-driven techniques. The technology enables instantaneous communications between devices and the environment.
A new study from the University of Birmingham suggests that epidemic modeling can significantly overestimate infection numbers due to ignoring network structures. The researchers found that heterogeneous networks can lead to smaller epidemic waves and fewer infections, contradicting standard 'random mixing' models.
A new AI system developed by the University of Technology Sydney can rapidly detect COVID-19 from chest X-rays with high accuracy. The Custom Convolutional Neural Network (Custom-CNN) model streamlines the detection process, providing a faster and more accurate diagnosis.
A novel human brain organoid model generates all major cell types of the cerebellum, including functional Purkinje neurons. This breakthrough provides a new way to explore cerebellar development and disorders, advancing therapeutic interventions.
Scientists have created a self-organizing neuromuscular junction model from human pluripotent stem cells to study complex neuromuscular diseases. The 2D and 3D cultures mimic the physiological situation, allowing researchers to perform high-throughput drug screening for novel treatments.
A recent study using AI to analyze registry data on people's residence, education, income, health, and working conditions can predict life events such as personality and time of death. The model outperforms other advanced neural networks and provides precise answers despite ethical concerns about sensitive data and bias.
A new study from MIT shows that computational models trained on auditory tasks display an internal organization similar to the human auditory cortex. Models trained on diverse tasks and background noise more closely mimic brain activation patterns.
Researchers at West Virginia University are using artificial intelligence to analyze habanero peppers and develop new methods for predicting genetic traits. The goal is to improve crop yields and prevent genetic diseases, with potential applications in human health.
Researchers use AI to develop dynamic modeling of brain graphs, capturing dynamics in continuous time for more accurate predictions and personalized treatment of brain diseases. The project aims to track disease development in individual patients and identify biomarkers associated with brain disorders.
A research team at City University of Hong Kong has developed a novel performance evaluation method called Information Exchange Surrogate Approximation (IESA) to calculate blocking probabilities in queueing systems with overflow. IESA provides ways to allocate limited resources better, enhancing resource allocation and capacity plannin...
Researchers used AI-selected natural images and synthetic images to probe visual processing areas of the brain, finding that predicted maximal activator images significantly activated targeted areas. The study suggests individualized models for each subject can improve understanding of visual system organization across populations.
Researchers have developed new tools to assess disease progression of Alzheimer's disease in animal models, providing a translational approach for studying the disease. The tools use neuroimaging and network modeling techniques to analyze metabolic changes in the brain, confirming previous clinical findings.
Researchers developed three diffractive deep neural networks using orbital angular momentum to recognize objects in images, achieving accuracy comparable to wavelength and polarization-based models. The technology has potential for real-time processing applications like image recognition and data-intensive tasks.
The paper proposes an adaptive representation model for geoscience knowledge graphs that can efficiently represent complex spatiotemporal features and relationships. The model uses a unified spatiotemporal ontology to automatically adopt an adaptive tuple structure based on the association of spatiotemporal information.
A new framework demonstrates that proportionately more multi-homing consumers lead to significant efficiency gains when integrating two business platforms. However, this trend also creates higher barriers to entry for new platform firms and may require policy guidance to mitigate potential harms of platform mergers.
Researchers developed a deep learning model that can identify previously unknown quasicrystalline phases in multiphase crystalline samples. The model achieved a prediction accuracy of over 92% and successfully detected an unknown phase in Al-Si-Ru alloys.
A new MIT study proposes a theoretical model that helps explain how cells maintain the memory of their cell type despite losing chemical modifications during DNA replication. The research team suggests that the 3D folding pattern of the genome determines which parts will be marked by these chemical modifications.
A new data-driven framework uses machine learning to identify keystone microbial species in human gut microbiomes, varying across communities. The study found that keystone species have context-dependent essentiality and can aid digestion, breaking down complex starches.
Researchers at Purdue University developed a new tool to visualize neural network decisions, making it easier to identify errors in image recognition. The tool uses graph-topological data analysis to provide a bird's-eye view of all images in a database, revealing areas where the network struggles to distinguish between classifications.
A new algorithm from the University of Surrey models distributed electricity networks, finding that local renewable energy generation is generally more efficient than central storage or export. The study suggests that considering local constraints and factors like energy prices and subsidies can help design the most efficient local grids.
A recent study published in the Proceedings of the National Academy of Sciences found that AI's deep convolutional neural networks can identify faces but struggle to capture other important information like emotional state and trustworthiness. Brain activity scans revealed a weak correlation between AI's codes and human brain represent...
Researchers found that drones could improve access to automated external defibrillators (AEDs), a portable device for treating out-of-hospital cardiac arrests. The simulation model suggests that drone delivery systems could reduce response times in urban and rural areas, with greater improvements expected in rural areas.
Researchers are combining biology, physics, computer science, and engineering to design electric circuits that mimic the brain's adaptive behavior. The goal is to create a more efficient AI application that can learn from history and adapt without significant energy consumption.
Researchers developed an AI model to optimize network allocation, saving bandwidth and reducing computational cost. The model can be adapted for various scenarios, including drone battery conservation and remote surgery.
Researchers found self-supervised models generate activity patterns similar to mammalian brains, suggesting an organizing principle. The models learn representations of the physical world to make accurate predictions, potentially unlocking human-labeled data limitations.
A new project aims to help robots assess risks and make autonomous decisions. The research focuses on quantifying ambiguity in robot perception to improve safety and efficiency.
A new method called TWC-Swin effectively restores holographic images even under low spatial coherence and arbitrary turbulence, surpassing traditional convolutional network-based methods. The study demonstrates strong generalization capabilities, extending its application to unseen scenes.
Researchers from Austria and France join forces to unravel the secrets of gene regulation during mammalian development using stem cell-derived 3D culture models. The project aims to understand how key molecular events influence gene transcription and regulation over hours and days.
A new neighborhood-based care model has been shown to be effective in treating hepatitis C among injection drug users and those experiencing homelessness. The study found that 92% of participants had undetectable levels of the virus after treatment, with 84% achieving sustained virologic response.
Researchers at MIT found that similarity-focused generative AI models falter when tasked with designing new products, highlighting the need to prioritize innovation in engineering tasks. By adjusting training objectives and metrics, AI can be an effective 'co-pilot' for engineers, enabling faster creation of innovative products.
A new analysis method can map individual company-level connections, revealing up to 13 billion supply connections worldwide. This data could reduce tax evasion (€130 billion) and improve climate and human rights compliance by creating a global supply network map.
Researchers from GIST propose a novel approach to mitigate overfitting in pretrained models used for voice pathology detection, achieving 12.36% and 15.38% improvement in recall using contrastive learning.
Researchers found that deep neural networks often respond the same way to images with no resemblance to the target, generating unnatural signals. The models develop unique invariances that are different from human perceptual systems, causing them to perceive pairs of stimuli as similar despite their differences.
Researchers at Linköping University developed an AI-based method applicable to various medical and biological issues, accurately estimating people's chronological age and determining smoking status. The models identify previously known epigenetic markers used in other models, but also new markers associated with conditions.
The study developed a highly accurate AI model for fully automated cancer detection, including small and difficult-to-detect tumors. The model could detect visually imperceptible cancer from normal-appearing pancreases substantially early before clinical diagnosis, with a median of 438 days.
Researchers developed a model demonstrating that chasing interactions can induce dynamical patterns in bacterial organization. The structure becomes visible on a higher level, without requiring adhesion or alignment.
Post-Acute Sequelae of COVID-19 research aims to track long-term health symptoms in survivors. A $3.7 million grant will support the development of self-supervised deep learning technologies to recognize post-COVID lung progression phenotypes.