Articles tagged with Artificial Intelligence
Researchers developed a model to predict household energy costs with over 74% accuracy by analyzing passive design characteristics from Google Street View images. The model can help identify neighborhoods vulnerable to energy burden and pave the way for smart, sustainable cities.
A breakthrough metamaterial has been developed to enable real-time shape and property control, surpassing existing materials' limitations. The material's adaptable capabilities make it suitable for applications in robotics and other fields requiring flexibility.
A new research review at ECCMID 2024 explores the use of AI in infection control, including ensuring correct mask wear and intelligent cleaning. The key to successful AI use is identifying problems and offloading repetitive tasks to AI systems.
Scientists studied thousands of public photos to track humpback whale populations, finding they're responding to food availability changes affected by climate change. The population declined by 7,000 individuals from 2012 to 2021 due to a massive marine heatwave and reduced prey.
Researchers warn that AI tools used in radiology contribute to greenhouse gas emissions through energy consumption and data storage. To minimize these impacts, experts recommend sharing resources, selecting sustainable hardware, and using data compression techniques.
A recent UConn study assessed the accuracy of ChatGPT's exercise recommendations, revealing that the AI provided only 41% of the expected content in a gold-standard recommendation. The chatbot demonstrated high accuracy for general guidelines but failed to provide key elements such as frequency, intensity, and type of physical activity.
Research suggests AI can identify key risk factors for outbreaks, pinpointing specific
Researchers developed an AI-driven lab called Fast-Cat that uses artificial intelligence to provide in-depth analyses of catalytic reactions. The tool conducts high-temperature gas-liquid reactions and analyzes results to determine how different variables affect the outcome of each experiment.
A new smartphone application using AI and facial-image processing software detects the onset of depression before symptoms appear. MoodCapture captures facial expressions and surroundings, then evaluates them for clinical cues associated with depression.
A Tohoku University team studied deep learning models for diagnosing drowning and found discrepancies between model results and medical experts' observations. The study highlights the need for new training methods that align AI model internals with human expertise.
PandaOmics uses advanced AI algorithms to process vast quantities of diverse data, performing gene and pathway analysis and target predictions. The platform has been extensively validated in multiple therapeutic areas, including oncology, inflammation, and immunology.
A new AI system is being developed to provide real-time feedback and instruction for surgical trainees, addressing the need for constant supervision in laparoscopic surgery. The system uses computer vision model to detect the student's actions and provides immediate assessment and training direction.
Explainable AI methods have been developed to make audio models more interpretable and transparent. Researchers categorize existing audio XAI methods into two groups: general methods and audio-specific methods, offering new possibilities for improving the trustworthiness of AI decision-making in audio tasks.
Researchers identify five grand challenges in biomedical engineering to address social needs, existing gaps, and technological limitations. The Convergence Revolution and Fourth Industrial Revolution are expected to shape the future of medicine, emphasizing interdisciplinary collaborations and next-generation training.
The USC team created a low-cost, customizable learning kit for students to build their own 'robot friend' using the Blossom robot. The three-part module provides hands-on experience and instruction on various AI aspects, including robotics, machine learning, and software engineering.
A new AI-powered system has been shown to detect 40% more polyps during colonoscopies, a crucial step in preventing colorectal cancer. This breakthrough technology uses artificial intelligence to examine every pixel in the image and alert doctors to potential issues, allowing for more accurate diagnoses.
A team of pediatric clinicians developed 45 multilingual videos to educate providers on pediatric care topics, using generative AI and synthetic voice avatars. The videos were recorded in English and translated into Spanish, emphasizing Latin American variations.
A new study published in JAMA Ophthalmology suggests that artificial intelligence can match or exceed the expertise of seasoned ophthalmic specialists in diagnosing and treating patients with glaucoma and retina disorders. The AI system, GPT-4, demonstrated superior performance in response to glaucoma questions and case-management advice.
A new process using artificial intelligence (AI) predicts carbon cycles in agroecosystems, surpassing traditional models in accuracy and speed. This breakthrough enables fair and accurate compensation for farmers, fostering trust in carbon markets and promoting sustainable practices.
Researchers have developed an AI-driven method to detect possible anomalies beneath the surface of wind turbine blades using patented radar technology. The non-destructive approach supports agile data acquisition and analysis, enabling faster detection of manufacturing defects and improved overall quality assurance.
Researchers used AI tool AlphaFold to predict and model over 1,400 protein-protein interactions essential for bacterial survival. The study reveals previously unknown details and offers potential targets for developing new antibiotics.
A Princeton University team developed an AI model that can forecast potential plasma instabilities up to 300 milliseconds in advance, allowing for real-time adjustments to avoid reaction-ending escapes. The model uses past experimental data and demonstrates a promising approach to solving a broad range of plasma instabilities.
A new study by EPFL researchers Antonia Sclocchi and Matthieu Wyart examines the impact of changing batch size and learning rate on stochastic gradient descent. They identify three regimes that affect AI learning differently, from exploratory to efficient approaches.
A study published in PNAS Nexus found that AI-generated propaganda was almost as effective as real propaganda in persuading US adults. The research used a large language model to generate new articles based on existing propaganda, with 43.5% of respondents agreeing with the claims after reading the AI-written articles.
Researchers at Simon Fraser University have developed an AI approach to intrinsic decomposition, separating images into lighting effects and true colours of objects. This innovation enables a range of applications, including image compositing and augmented reality, making it accessible to content creators and post-production artists.
A recent study found that parents and clinicians are in favor of a mobile app to help manage child cancer pain at home. The app aims to ease the burden on caregivers and provide personalized pain management support. Recommendations for the app include improved accessibility, language options, and gamification features to involve children.
The NSF director highlights the benefits of AI in scientific research, from accelerating discovery to automating routine tasks. He emphasizes the need for safe and accessible AI, calling for a diverse AI ecosystem that incorporates broad perspectives and problem-solving approaches.
The Heidelberg Laureate Forum Foundation offers limited international journalist travel grants to report on the event. Journalists can apply until April 28, 2024, and will cover travel and visa costs, as well as room and board during their stay in Heidelberg.
Researchers at North Carolina State University are developing a suite of performance metrics to standardize the evaluation of self-driving labs in chemistry and materials science. These metrics aim to compare different lab technologies and identify areas for improvement, ultimately advancing the field and accelerating discovery.
A KAUST research team has developed a machine-learning approach that balances privacy preservation and model performance using ensemble privacy-preserving algorithms. The approach, called PPML-Omics, achieves better model performance while keeping the same level of privacy protection compared to previous methods.
Researchers from Argonne National Laboratory and the University of Illinois Urbana-Champaign used generative AI to quickly assemble over 120,000 new MOF candidates for carbon capture. The approach combines AI with high-throughput screening, molecular dynamics simulations and theory-based design to identify optimal materials.
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.
Researchers analyzed over 1,000 studies to identify design principles for creating empathic conversational agents. The study reveals the importance of emotional intelligence and adaptive feedback in improving learning outcomes.
Researchers used data from 9,300 miles of Greek roads to develop a machine-learning model predicting crash sites. The model identified key features such as abrupt speed limit changes and incomplete lane markings as predictors of crashes. The study's findings have implications for improving road safety globally.
A study found that customers are more willing to provide written compliments and complaints using AI-powered tablets rather than leaving videos. The researchers suggest that methods requiring 'low self-disclosure' would work better for customer feedback.
Researchers at UC Santa Cruz find that popular link prediction metrics are flawed and do not accurately measure algorithm performance. They recommend using a new metric, VCMPR, to benchmark link prediction tasks and highlight the importance of accurate metrics in machine learning decision-making.
A new depth from focus/defocus approach, DDFS, combines model-based and learning-based strategies to achieve notable improvements in performance and applicability. The proposed method outperformed state-of-the-art methods in various metrics for several image datasets.
A new paper argues that LLMs can interpret and analyze neuroscientific data, unlocking new insights and potential treatments. Lead author Danilo Bzdok suggests that scientists may not always fully understand the mechanism behind biological processes discovered by LLMs.
A new tumor spatial mapping tool will help clinicians assess the aggressiveness of cancer and personalize treatment. The tool maps the functional state of oncoproteins onto tumour images, allowing for more accurate diagnosis and informed treatment decisions.
Researchers developed sensors using aerogels to detect formaldehyde, a common indoor air pollutant, with real-time detection capabilities. The sensors require minimal power and can distinguish between different gases.
Researchers developed an AI model to detect low blood sugar levels in diabetics while driving, improving safety and preventing accidents. The model uses driving behavior and head/gaze motion data, showing promise for future self-driving cars.
A clinical trial found that patients who received care recommendations from an AI-based system had fewer recurrent strokes, heart attacks, or vascular death within three months. The study suggests that this technology can improve patient outcomes and aid neurologists in making informed decisions.
Researchers developed an AI algorithm to analyze brain scans and clinical data, predicting treatment response with 8-week accuracy.
Assistant Professor Santiago Segarra at Rice University has won the NSF CAREER Award to develop a new approach for AI-powered climate prediction by leveraging structural properties in real-world data. The research aims to create more effective learning algorithms for structured domains.
Researchers at Forschungszentrum Jülich have programmed an AI that can recognize patterns in complex data sets and formulate them in a physical theory. The AI uses physics for machine learning to simplify complex interactions and build up a new theory.
The JGU Center for Lifelong Learning is developing personalized AI-based learning experiences for adult learners, aiming to improve motivation and learning outcomes. The project will also discuss the challenges of AI technologies in teaching modern foreign languages.
A new study uses GPT-3 to simplify chemical analysis, achieving accuracy surpassing state-of-the-art models. The approach fine-tunes the language model with curated Q&As, enabling easy and fast discovery in low-data chemistry.
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.
Researchers from Indiana University School of Medicine examined the use of ChatGPT in writing scientific review articles. They found that while AI-assisted approaches decreased writing time, they required more extensive fact-checking and produced incorrect references up to 70% of the time.
The multidisciplinary team, led by Nihar Mahapatra, is developing sustainable and impactful accessible voice AI solutions for people who stutter. The goal is to provide high-impact deliverables that will be sustainable beyond their support.
Researchers at ISTA have discovered the composition of poxviral cores, a key factor in their infectivity. The study's findings could lead to the development of new therapeutics targeting the viral core.
Smarter eco-cities harness real-time data to address environmental challenges and implement innovative approaches for ecological conservation. However, high energy demands associated with AI and AIoT applications pose a challenge to attaining environmental objectives.
Researchers from King's College London developed an AI model to predict brain cancer patient outcomes after radiotherapy. The model showed improved performances when trained on 10,000s of brain MRI scans and can give clinicians accurate predictions with a single routine scan, allowing for early treatment decisions.
A neural network trained on a photographic database of olive fruit endocarps can identify olive varieties with high accuracy. The OliVaR tool automates the traditional morphological classification process, allowing growers to quickly identify olive tree varieties and advancing general knowledge of all existing olive varieties.
A new study published in JAMA highlights the need for federal regulations that prioritize patient outcomes when it comes to AI-powered tools in healthcare. The researchers argue that AI products in healthcare do not universally undergo evaluation for potential harm or benefit to patients, and propose a revision to the White House Execu...
AI-based personalized medicine is developing rapidly, enabling tailored therapies for individual patients. However, existing regulatory frameworks create significant challenges for approving these novel treatments.
A new machine learning approach combines computer vision with deep-learning algorithms to pinpoint problem areas in concrete structures. The system enables efficient identification and inspection of cracks using autonomous robots, reducing the overall inspection workload.
Lehigh University researcher Dominic DiFranzo is developing an AI-powered platform, Social Media TestDrive, to equip youth with skills to navigate social media safely. The platform uses conversational AI tools to provide instant feedback and interactive simulations, empowering students to be 'upstanders' against cyberbullying.
Researchers have developed an AI-driven test that accurately diagnoses ovarian cancer in women clinically classified as normal, improving detection of early-stage disease. The test uses machine learning and blood metabolite information to assign a probability of disease presence or absence, offering a more clinically informative approach.
A new study proposes a data-sharing solution based on distributed fountain coding to improve the reliability of transmission in complex vehicle-connected environments. The proposed method allows receivers to accept encoded packets from different vehicles, reducing decoding failure caused by network anomalies.