Articles tagged with Machine Learning
Researchers at New York University developed a novel learning procedure called Meta-learning for Compositionality (MLC) that enables neural networks to make compositional generalizations. MLC outperforms existing approaches and is on par with, and in some cases better than, human performance.
A new property evaluation method for nanoparticles' shape anisotropy has been developed using deep learning, achieving classification accuracy of approximately 80% on single particle basis. This breakthrough solves a long-standing issue in nanoparticle evaluation dating back to Einstein's time.
NTU Singapore has expanded its research collaborations with French partners to push the boundaries of science. The university has inked six new partnerships and renewed existing collaborations across various fields, including quantum physics, nuclear energy, and sustainability.
A new AI-based tool estimates patient-specific cancer survival rates using machine learning, offering personalized prognosis for patients with breast, thyroid, and pancreatic cancers. The Cancer Survival Calculator differs from existing estimators by incorporating multiple factors beyond disease stage.
Researchers used Fourier-transform infrared spectroscopy and machine learning to predict adsorption capacity of pharmaceuticals and personal care products on long-term aged microplastics. The study successfully captured the complexity of the system with up to 98% accuracy, providing new insights into the interactions between microplast...
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 proposed AI-centric medical curriculum aims to educate future healthcare practitioners in digital technology, with a focus on technical concepts, validation, ethics, and appraisal. The curriculum caters to varying student levels, from consumers to developers, promoting interprofessional collaboration and adaptable learning.
Lehigh University researchers have developed a technique using machine learning and advanced spectroscopy to characterize waste feedstocks for gasification-produced hydrogen. This process has the potential to eliminate hazards associated with stored coal waste and reclaim valuable resources, while also emitting fewer pollutants than tr...
Lerrel Pinto aims to create robots that can perform a variety of tasks using algorithms and data collection methods. He seeks to address the trade-off between dexterity and generality in robotics.
Scientists designed an mRNA nanovaccine using machine learning to overcome delivery barriers, promoting strong immune responses and activating the STING pathway to kill tumor cells. The therapeutic strategy demonstrated stronger anti-tumor effects in melanoma and colorectal cancer models.
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 from Mindgard and Lancaster University exposed vulnerabilities in major Large Language Models (LLMs) that can be copied for as little as $50. This 'model leeching' attack allows attackers to gain insights into how LLMs work, enabling them to launch targeted attacks with an increased success rate.
Researchers developed a new method to estimate gradients and derivatives on quantum computers, enabling faster computations. This technique can be applied to various fields such as cryptography, optimization, and materials science.
Researchers developed an AI-based method to estimate BMD from plain X-ray images using a hierarchical learning framework. The approach showed high performance and reliability in estimating BMD, with correlation coefficients of 0.88 and 0.92 compared to DXA and QCT.
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.
A research team led by HKUST developed an AI-powered model to predict glioma patients' prognosis and identify early predictors of tumor evolution under therapy. The model, CELLO2, uses genomic and transcriptomic data from 544 glioma patients to accurately predict treatment-induced hypermutation and grade progression.
A new generative model named scPoli enables multi-scale representations of cells and samples, facilitating the integration of high-quality large-scale datasets for novel biological insights and disease understanding. This model accelerates atlas building and usage, ultimately accelerating disease understanding and therapy development.
Researchers found that using predictive models in healthcare can alter relationships between patient data and outcomes, leading to further degradation. Implementing a system to track individuals impacted by machine learning predictions is crucial to maintaining model performance.
A team of New York University computer scientists created a neural network that can explain how it reaches its predictions, shedding light on the intricacies of RNA splicing. The breakthrough reveals how a small, hairpin-like structure in RNA can decrease splicing and provides new insights into the transfer of genomic information.
The Army Research Laboratory has chosen Texas A&M University for its High-Throughput Materials Discovery for Extreme Environments Center (HTMDEC). The center aims to develop novel materials for extreme conditions, reducing experimentation costs and duration. By leveraging machine learning, physics-based simulations, and collaboration, ...
Researchers at the University of Texas at Austin developed an AI algorithm that accurately predicted 14 earthquakes within about 200 miles of their location and strength, with only one false warning. The system detected statistical bumps in real-time seismic data and paired them with previous earthquakes to make predictions.
A new machine learning approach called CellOT allows for precise modeling and prediction of cell changes and drug effects. This enables personalized cancer treatments by identifying the traits of resistant cells, paving the way for more effective therapies.
Researchers from Imperial College London and the University of Nottingham used machine learning to identify 'atomic shapes' that form basic pieces of geometry in higher dimensions. The findings reveal unexpected patterns in these shapes and demonstrate the potential for machine learning to accelerate mathematical discoveries.
Researchers discovered that monk parakeets possess a unique tone of voice, known as a voice print, similar to humans. This finding raises the possibility that other vocally flexible species may also have a voice print.
Researchers from SUTD successfully applied reinforcement learning to a video game problem, creating complex movement designs that outperformed top human players. The study's findings have the potential to impact robotics and automation, ushering in a new era of movement design.
Researchers discovered that users' prior beliefs about an AI chatbot's motives significantly impact their interactions with the agent. Priming users to believe certain things about the AI's empathy, neutrality, or manipulation influences their perception of its trustworthiness and effectiveness.
A new AI method leverages causal relationships in genome regulation to efficiently identify optimal genetic perturbations for cellular reprogramming. The technique reduces experimental costs by prioritizing the most informative interventions, leading to faster convergence and more effective results.
Researchers from FAU's College of Engineering and Computer Science employ a computer-vision deep learning technique to analyze wall-bounded turbulent flows. They successfully identify the sources of extreme events in a data-driven manner, providing new insights into non-linear relationships in fluid dynamics simulations.
A new study found that ChatGPT-4 can generate highly accurate personal narratives based on stream-of-consciousness thoughts and demographic details. The AI model was used in conjunction with therapists to guide patients toward healthier thoughts and behaviors, suggesting a potential tool for improving therapeutic approaches.
Ghent University's research team explores active machine learning to revolutionize chemical engineering. The technology offers greater flexibility and performance compared to traditional design of experiment algorithms.
Researchers are working on EdgeRIC: Real-time radio access network intelligent control, aiming to enhance wireless communication and machine learning. The project involves developing intelligent systems that can adapt to changing conditions, prioritizing users and acquiring lower latency for applications.
A new modeling method powered by interconnected processors removed human bias from the debate over dinosaurs' demise. The study suggests that the outpouring of climate-altering gases from the Deccan Traps alone could have been sufficient to trigger global extinction, consistent with volcanic eruptions contributing to the mass extinction.
Researchers have identified a major weakness in reservoir computing, a powerful machine learning tool used to model complex dynamic systems. The tool requires a lengthy warm-up time and relies on key information about the system being predicted being built in, making it challenging to accurately predict chaotic behaviors.
Insilico Medicine has identified 9 potential dual-purpose targets against aging and 14 major age-related diseases using Microsoft BioGPT. The proposed genes include CCR5 and PTH, which have not been previously correlated to the aging process.
The University of Trento coordinates a European research network, Elias, aiming to establish Europe as a leader in AI research. The project focuses on developing new computational systems and models for sustainable innovation, social cohesion, and reliable AI.
A new clinical and research partnership has created an AI model that can predict whether cancerous tissue has been fully removed from the body during breast cancer surgery. The model performed as well as humans in identifying positive margins, especially in patients with higher breast density.
A recent study published in the Journal of Chemical Information and Modeling presents a significant breakthrough in accelerating giga-scale virtual screens using machine learning. The researchers successfully reduced processing time by 10-fold for 1.56 billion drug-like molecules, identifying top-scoring compounds in under ten days.
Researchers used machine learning to expand and accelerate work on 'atomic shapes,' fundamental pieces of geometry in higher dimensions. The breakthrough identifies shapes and their properties, such as dimension, accelerating new insights across Pure Mathematics.
Researchers at the University of Zurich developed PlantServation, a method that enables scientists to observe plants with great precision using AI and machine learning. The technique allows for the analysis of millions of images taken from various weather conditions, providing insights into how plants respond to environmental factors.
The UW team's system uses self-deploying microphones to divide rooms into speech zones and track the positions of individual speakers. The robots disperse as far from each other as possible, allowing for greater sound control and isolating specific areas or separating simultaneous conversations.
A new study uses AI to predict mortality rates in patients with hip fractures, identifying key biomarkers for poor outcomes. The model achieved accurate predictions using basic blood and lab test data as well as demographic information.
Researchers from University of Cambridge and Cornell University have developed a method to build machine learning models that can understand complex equations using far less training data. This breakthrough enables the construction of more time- and cost-efficient models for physics, engineering, and climate modeling applications.
The project aims to assess the operational resilience of microgrids on DoD installations and ships, using new operational resilience indexes developed by Lehigh University researcher Javad Khazaei. The team will develop a dashboard to monitor resilience indexes in real-time, providing recommendations for improving the systems.
A new study found that AI-tutored students caused less damage to healthy tissues and improved safety measures, but also showed negative outcomes such as reduced efficiency and speed. Human instructors are essential to promote both safety and efficiency in neurosurgical training.
A research team at Toyohashi University of Technology has developed a technique to create training data for robots that estimate the state of users using machine learning. The method uses a human body link model without requiring movement analysis, enabling care robots to assist elderly with reduced burden and improved safety.
A new study introduces the concept of a 'noise barometer,' measuring biological age based on epigenetic changes. The researchers found that specific cytosines show increased variability with age, indicating a biomarker for aging and disease. This approach provides a novel perspective on aging research.
Researchers found that AI/ML based programs can successfully detect PCOS with an accuracy of 80-90%, making it a promising tool for early diagnosis and reducing the burden on patients. The study suggests integrating large population-based studies with electronic health datasets to identify sensitive diagnostic biomarkers.
A UMass Amherst neuroscientist is mapping the brain of a sea slug to study how neurons are added to functional neural circuits, shedding light on how this process contributes to neurological conditions. The project aims to provide an unprecedented look at brain development and potentially inform human brain development.
A study by researchers at the University of Illinois Urbana-Champaign found that YouTube's recommendation system did not promote anti-vaccine content during the COVID-19 pandemic. The study analyzed over 27,000 video recommendations and found that users were directed to longer, more popular health-related content.
Using computational topology, Brown researchers have developed an algorithm that profiles shapes and spatial patterns in embryos, enabling the study of how cells assemble into tissue-like architectures. The new approach uses persistence images to rapidly compare large datasets, reducing computation time from hours to seconds.
A new study from UTHealth Houston finds that AI-powered algorithm can improve detection rates of unruptured cerebral aneurysms. The study used a machine learning algorithm to analyze CT angiograms and identified 36 true aneurysms, with 24 previously not referred for follow-up.
A new method of analyzing nanoscale X-ray movies reveals unprecedented insights into how lithium-ion batteries store and release charge. The study suggests ways to improve the efficiency of billions of nanoparticles in electrode materials, potentially leading to faster-charging batteries.
Researchers developed an AI foundation model for eye care that can identify sight-threatening eye diseases and predict general health. The model, RETFound, was trained on millions of eye scans from the NHS and outperforms existing state-of-the-art AI systems across complex clinical tasks.
Researchers created a machine-learning algorithm using the largest data set of digital biomarkers for hot flashes. The predictive system delivers cooling to mitigate or fully alleviate hot flashes using Embr Wave's wearable device.
Researchers propose a reinforcement learning-based approach to optimize multi-impulse linear rendezvous trajectories, achieving faster computation times and improved fuel efficiency compared to traditional numerical optimization methods. The algorithm uses an actor-critic architecture and advantage-weighted learning to accelerate train...
A new study published in JAMA found that adults over 60 who spend more than 10 hours a day engaging in sedentary behaviors like sitting are at increased risk of developing dementia. The study used wearable accelerometers to track physical activity and found that the total time spent sedentary each day was a significant predictor of dem...
Australian researchers analyzed over 1,300 Golden staph strains, linking specific genes to antibiotic resistance and the bacteria's ability to linger in the bloodstream. The study highlights the diagnostic power of integrating clinical and genomic data to develop targeted solutions for deadly superbug infections.
A Los Alamos-developed machine learning algorithm successfully processed massive data sets exceeding a computer's available memory. The algorithm divides data into manageable batches to prevent hardware bottlenecks, enabling efficient processing of large-scale applications in various fields.
A recent study has successfully predicted potential drug outcomes and side effects by analyzing the discrepancy in gene perturbation effects between cells and humans. Researchers used machine learning to forecast drug approvals, improving reliability over conventional methods that only consider chemical properties.
A new AI tool assesses motor performance using finger taps, providing rapid and standardized ratings, comparable to expert neurologists. The test outperformed primary care physicians in accuracy, opening doors to evaluating other movement disorders.