Articles tagged with Artificial Intelligence
Scientists at ETH Zurich used AI to analyze data from 1,380 borylation reactions and predict optimal synthesis methods for new drugs. The model was tested on six known drug molecules and showed a success rate of five out of six cases.
Researchers developed an AI tool that can identify never-smokers at high risk for lung cancer based on their chest X-ray images. The study found that 28% of non-smokers were deemed high risk by the model, and these patients had a 2.1 times greater risk of developing lung cancer compared to low-risk individuals.
Research analyzing over 1.4 million Americans finds significant disparities in automation job displacement risk across race and gender. The study highlights the importance of education in reducing automation risks and addressing existing racial and gender disparities.
Researchers develop AI methods to analyze photoluminescence data and identify factors influencing coating quality. The findings provide a blueprint for improving production processes and boosting the efficiency of highly efficient solar cells.
Virginia Tech researchers analyzed partisan media sentiment toward AI and found that liberal-leaning media tend to have a more negative tone than conservative media. The study suggests that this opposition can be attributed to concerns over AI amplifying existing social biases, such as racial and income disparities.
Artificial intelligence tools can generate convincing texts, images, voices, and videos, making it difficult to distinguish misinformation from genuine content. To combat this, experts suggest being more alert when consuming online content and relying on trusted sources.
Researchers at Tufts University have created hybrid transistors using silk proteins that can detect changes in humidity, oxygenation levels, and glucose. The transistors have the potential to enable integrated circuits that train themselves and respond to environmental signals.
A new AI program created by researchers at UF and NVIDIA can generate medical records so well that human physicians couldn't tell the difference. The GatorTronGPT model uses a large language model to mimic natural human language, overcoming hurdles such as protecting patients' privacy and being highly technical.
A computer simulation by Nagoya University researchers found that human behavior, such as lockdowns and isolation measures, influenced the evolution of new COVID-19 strains. The study discovered that SARS-CoV-2 variants with higher peak viral loads were more successful at spreading, but also had shorter infection durations.
Scientists at the University of Copenhagen and University of Victoria have developed an AI formula to predict rogue waves, which can split apart ships and damage oil rigs. The new knowledge can make shipping safer by identifying the likelihood of being struck by a monster wave at sea.
A study published in Facial Plastic Surgery & Aesthetic Medicine found that ChatGPT outperformed expert plastic surgeons in answering preoperative and postoperative patient questions. The AI tool received significantly higher ratings for completeness and overall quality.
A recent study by Flinders University researchers found that Generative AI can rapidly create convincing disinformation on health topics, including fake videos and articles.
Researchers suggest a new evaluation framework to assess AI reasoning abilities, comprising psychological experiments, self-reflection, and source code analysis. This approach aims to determine if AI systems genuinely reason like humans.
Scientists at PNNL introduced a new way to evaluate AI system recommendations by incorporating human experts' insights. Human expertise improved the accuracy of predictions and boosted confidence scores, indicating better decision-making capabilities for machine learning systems.
A recent study found that during stringent COVID-19 periods, online searches surged in the health and daily life category. As government policies relaxed, searches shifted towards duty-free and travel-related products. The research team applied PCA to Big Data of internet search activity volume data from NAVER DataLab platform.
A novel technology to manage demands on mobile networks from multiple users has been developed by University of Leicester computer scientists. The study found a 10% power consumption reduction compared to existing technologies, with faster device selection and less resource allocation.
A paper by Anthony Chemero explains how AI thinking differs from human thinking, highlighting the limitations of large language models trained on biased data. Despite generating impressive text, these models can make up facts and produce biased outputs due to their lack of embodiment and understanding of context.
The Python code library snnTorch, developed by UC Santa Cruz's Jason Eshraghian, has surpassed 100,000 downloads and is used in various projects. A new paper published in the Proceedings of the IEEE documents the library and offers a candid educational resource for students and programmers interested in brain-inspired AI.
A University of Louisville law professor is creating a generative AI toolkit to aid legal writing instruction, providing resources for professors to incorporate the technology into their curricula. The open-source materials will enable instructors to customize their use of genAI and align teaching objectives with student outcomes.
A new deep-learning platform, EUGENe, simplifies data analysis for genomics researchers. The software can be adapted to various projects and reproduces results from existing studies.
A research team employed deep learning techniques to scrutinize dam operation patterns, achieving remarkable accuracy in forecasting dam water levels. The study demonstrates the potential of an artificial intelligence model trained on extensive big data to surpass conventional physical models.
A team of researchers has developed an atom-predicting model similar to the GPT models that support applications like ChatGPT. The new model focuses on small organic molecules with relevance to energy storage and conversion applications.
A novel robotic system developed by USC researchers can help clinicians accurately assess a patient's rehabilitation progress. The method generates an 'arm nonuse' metric using machine learning and a socially assistive robot to track how much a patient is using their weaker arm spontaneously.
A new deep learning AI tool called ECOGEN has been developed to generate lifelike bird sounds, enhancing the samples of underrepresented species. This allows for improved bird song classification accuracy and contributes to the conservation of endangered bird species.
Researchers developed a new 3D inkjet printing system that works with a wider range of materials, including slower-curing materials. The system utilizes computer vision to automatically scan the print surface and adjust the amount of resin deposited in real time.
Researchers developed a neural network called Senseiver that can reconstruct large systems from small amounts of sensor data using low-powered edge computing. The model has broad applications across industries, including climate modeling, self-driving cars, and medical monitoring.
Researchers at NC State University developed an autonomous system called SmartDope to synthesize 'best-in-class' materials for specific applications in hours or days. It uses a self-driving lab to manipulate variables, characterize optical properties, and update its understanding of the synthesis chemistry through machine learning.
A new study using twisted magnets as computational medium has made brain-inspired computing more adaptable, reducing energy use and potential carbon emissions. The research found that by applying magnetic fields and changing temperature, physical properties of the materials can be adapted to suit different machine-learning tasks.
This study investigates large language model (LLM) construction, optimization, and evaluation, highlighting the importance of open-source models and cost-saving methods. The authors also identify challenges faced by LLMs, including scarcity of datasets and model instability, and propose potential research directions.
Researchers found that smaller subsets of data can be just as effective in training AI models, reducing the need for massive computing power. The study suggests that information richness is more important than dataset size.
A new AI method combines satellite imaging and ecological analysis techniques to interpret large amounts of data from tumor tissue, providing insights into how cancer works. This approach aims to tailor cancer treatments to individual needs and avoid unnecessary side effects.
A new landmark study identifies 14 evolutionary traps that human societies are at risk of getting stuck in, including global climate tipping points, misaligned AI, and chemical pollution. To avoid these dead ends, the researchers emphasize the need for collective human agency and design settings where it can flourish.
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 at UC Berkeley introduce prediction-powered inference (PPI), a method to correct machine learning model output and provide valid confidence intervals. PPI allows scientists to incorporate AI predictions into their work without making assumptions about the model's limitations or data biases.
Yu Yang's NSF-funded research aims to reduce vehicle emissions and promote the use of electric bikes and scooters by developing socially informed traffic signal control systems. The project involves a three-pronged method that uses low-cost mobile air-quality sensing, spatial-temporal graph diffusion learning, and reinforcement learnin...
A recent study assessed ChatGPT's accuracy in identifying common allergy myths. The AI model correctly identified myths as true or false with an overall accuracy rate of 91%, with some myths being more accurate than others.
The project aims to develop a maturity model framework to outline essential capabilities for health systems to ensure trustworthy utilization of AI models. The framework will help identify strengths and weaknesses in procuring and deploying AI solutions, ultimately driving transformation of healthcare.
The UTSA MATRIX AI Consortium has received a $2 million grant to create new AI models that rapidly learn, adapt, and operate in uncertain conditions. The team aims to bridge the gap between human brain processing efficiency and current AI limitations, enabling more efficient and adaptive AI systems.
A new five-year research project combines AI, virology, and ecology to anticipate future SARS-CoV-2 strains that could pass between animals and people. The team will use artificial intelligence to predict variants and assess risk of spillover from people to wildlife.
The Ukraine War is a turning point in modern warfare, as new technologies like AI, drones, and cyberweapons are being used to devastating effect. Researchers like Jordan Richard Schoenherr warn that our understanding of warfare is outdated, and we need to rethink the role of sociotechnical systems in strategic thinking.
A study published in JMIR Medical Education found that GPT-4 can accurately diagnose and triage health conditions comparable to board-certified physicians. The model's performance does not vary by patient race or ethnicity, providing a promising tool for healthcare systems.
Researchers used sediment DNA to reconstruct a 100-year history of biodiversity, chemical pollution, and climate change levels in a Danish lake. The study found that pollutants like insecticides and increased temperatures had devastating effects on biodiversity, while suggesting some recovery over the last 20 years.
Researchers developed an AI system that can scan through college application essays to identify evidence of key personal traits, such as leadership and perseverance. The system aims to reduce algorithmic bias and provide more holistic admissions decisions.
The University of Würzburg's SONATE-2 nanosatellite is designed to test novel artificial intelligence (AI) hardware and software technologies in near-Earth space. The satellite aims to automatically detect anomalies on planets or asteroids, with the goal of improving planetary exploration and research.
Researchers developed personalized risk equations using AI to identify individuals at high risk of sudden cardiac death. The analysis found that AI was able to accurately predict sudden cardiac death in over one-fourth of all cases, highlighting the potential for AI to revolutionize prevention strategies.
A recent study published in Nature Communications validates MSIntuit CRC, an AI-driven digital pathology diagnostic, as a reliable pre-screening tool for colorectal cancer. The diagnostic accurately rules out nearly 50% of MSS patients while correctly classifying over 96% of MSI patients.
A study found that gastrointestinal and sleep issues may be connected to self-injury and aggression in adolescents diagnosed with profound autism. The researchers discovered a possible connection between these health issues and future challenging behaviors, predicting next-day behavior with over 80% accuracy.
Researchers have developed AI tools that can effectively detect heart valve disease and predict cardiovascular risk using digital stethoscopes. A study found that AI-powered digital stethoscopes predicted nearly 90% of valve disease diagnoses, offering a promising tool for transforming CVD care.
A team of researchers will develop a validated curriculum and assessment methods to increase ethical responsibility in the future cybersecurity workforce. The project aims to address social and ethical risks associated with AI technologies designed for security-related problems.
A new AI model trained by Cambridge researchers can classify 'hard-to-decarbonize' houses with high accuracy, enabling policymakers to prioritize improvement efforts. The model uses open-source data and can be adapted for use in countries with patchy datasets.
Researchers at the University of Sydney have developed a physical neural network that can learn and remember data in real-time, using nanowire networks to mimic brain-inspired learning and memory functions. The network achieved high accuracy in benchmark image recognition tasks and demonstrated its capacity for online learning.
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.
Professor Sang-hyun Park's research team developed AI technology that minimizes structural deformation in images while maintaining texture information from a new domain. This enables domain adaptation for deep learning models trained with generated images.
A team of scientists discovered two types of neurons in fruit flies and mice that enable them to identify distinct smells. With experience, these animals can learn to differentiate between very similar odors, a process that could improve machine-learning models and AI systems.
A new project will monitor how changing environmental conditions shape viral outbreaks in wild rodent populations to identify hotspots with high potential for spillover into people. The team will use metaviromics and AI to analyze data from wild rodents in the UK and Eastern Uganda.
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 used AI to identify 2 promising antigens as candidates for a gonorrhea vaccine, which accurately predicted reduction of bacterial populations. The antigens were tested in lab and animal models, showing efficacy in killing bacteria and decreasing bacterial burden.
A team of researchers from POSTECH successfully engineered a dual metalens capable of switching between different imaging modes using a single lens. This innovation enables fast mode-switching and acquisition of high-resolution images for applications such as bio-imaging and cellular reactions.
The DGIST research team developed an image translation model that can reduce biases in data despite the lack of information on underlying factors. The model achieved superior performance compared to existing methods on various biased datasets, including those with texture biases.
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.