Articles tagged with Software
A new data collection system has been developed to support diverse applications in the railway industry, including carbon emissions accounting and green cost-benefit analysis. The system enables efficient data collection and integrated management, leveraging metadata and data exchange schemas.
A new study uses reinforcement learning to develop an AI model that can hedge derivative contracts with significantly lower training data requirements. The model's adaptability and accuracy under realistic conditions outperformed benchmark systems.
The new platform, Arras Energy, simulates how all parts of an electric grid work together, making it easier to understand and apply the results. The tool can help utilities harden their distribution systems against extreme weather and integrate renewable energy sources.
Scientists at the University of Washington School of Medicine developed a novel protein design approach using AI, creating proteins that bind to challenging biomarkers with exceptionally high affinity and specificity. The breakthrough has implications for drug development, disease diagnosis, and environmental monitoring.
A NSF-funded project, MABLE, is developing a digital app using crowdsensing, AI, and robotics to empower individuals with responsive maps and turn-by-turn instructions. The app aims to improve accessibility and navigation for persons with visual or mobility impairments, such as those with low vision and wheelchair users.
Researchers developed a new solver algorithm for the MPAS-Ocean ocean circulation model, reducing run time by 45% and enabling semi-implicit stability. This allows for faster climate predictions and energy efficiency, as well as reduced computational power consumption.
A new algorithm, SIMSEF, significantly speeds up and simplifies the identification of chemicals in tissues, enabling researchers to confidently recognize and visualize their distribution in organs. This development has great implications for medical diagnosis and treatment outcomes.
The study reveals a diverse landscape of software as a medical device (SaMD) with notable progress in artificial intelligence/machine learning-based devices, accounting for 22% of all FDA-approved SaMDs. Established companies focus on incremental innovations, while new entrants and start-ups drive disruptive innovation.
A new software algorithm intelligently combines vocals and instrumentals from different songs, mimicking professional music composition techniques. The result is a unique blend of pleasing lyrics and exciting instrumentals that appeal to listeners, particularly hip-hop fans.
Professor Sebastian Erdweg aims to automatically incrementally compute software for significant energy savings. His new project AutoInc seeks to overcome fundamental limitations of existing approaches, promising enormous improvements in efficiency.
A new study reveals AI tools are more vulnerable than thought to targeted attacks that force AI systems to make bad decisions. Researchers developed a software called QuadAttac K to test for vulnerabilities in deep neural networks.
Researchers developed DIRFA, an AI-based program that generates realistic videos with facial animations synchronized to spoken audio, showcasing improvements over existing approaches. The tool has potential applications in healthcare, education, and entertainment, enhancing user experiences.
A team of researchers from Northwestern University and the University of Pittsburgh has received a $600K NSF Award to explore novel learning-enabled cyber-physical systems (LE-CPSs) for building flexibility into hardware. This could lead to extended device lifespan, reduced e-waste, and improved sustainability.
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.
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.
Venkat's project aims to reduce attack response time and protect programs from other attacks by developing a 'decoupled' security response system. His approach could help stop emerging zero-day cyberattacks within 24-48 hours, reducing time and dollar costs of frequent patching.
A Lancaster University academic argues that AI and algorithms contribute to polarization, radicalism, and political violence, posing a threat to national security. The paper examines how AI has been securitized throughout its history, highlighting the need for better understanding and management of its risks.
A recent study has shown that Apple's Safari browser is still vulnerable to Spectre attacks, allowing hackers to access sensitive data such as passwords and emails. Researchers from Ruhr-University Bochum found that attackers can bypass security measures in the browser to intercept and exploit user data.
Researchers developed a method combining sensor data with machine-learning algorithm to identify flaws in 3D-printed parts. The framework allows for statistically verified quality control, reducing the need for human involvement in manufacturing inspection.
A new study analyzes Twitter discussions on deepfakes related to the Russian invasion, highlighting negative reactions and positive sentiments towards deepfakes. The researchers found that some users' distrust in real videos increased after exposure to deepfakes.
A Rice University scientist used a mixed-reality headset with custom-designed software to measure and analyze forest floor vegetation in Tanzania's Udzungwa Mountains National Park. The study found that mammal diversity increases with forest floor habitat surface area, leading to a greater variety of ecological characteristics.
A new method using deep learning can provide as much information from brain images taken with CT as images captured with MRI, enhancing diagnostic support for conditions like dementia and other brain disorders. The software has been trained on 1,117 people and shows promise in diagnosing normal pressure hydrocephalus.
Serverless computing is a paradigm shift in cloud computing where cloud providers take over responsibility for complete server management. Two key principles define it: 'NoOps' and 'utilisation-based billing'. Research team from University of Würzburg explores the history, status and potential of serverless computing.
Researchers at UCLA found that GPT-3 performs well on analogy problems, similar to human undergraduates, but struggles with understanding physical space. The AI model's ability to mimic human reasoning is unclear, raising questions about its underlying cognitive processes.
Researchers developed a divide & conquer approach to leads-to model checking, mitigating the state explosion problem and improving performance. The technique, DCA2L2MC, divides the reachable state space into smaller sub-state spaces, making it feasible for large-scale systems.
A new open-source software, NMSM Pipeline, enables clinicians and engineers to create personalized computer models of patient movement to optimize treatment designs. The software uses physics-based models to predict and optimize functional outcomes for patients with various mobility impairments.
A University of Central Florida engineer is leading a $3.3 million ARPA-E funded project to develop simulation software for floating offshore wind turbines. The goal is to improve turbine design and increase their use as a renewable energy source. The software will be licensed or commercialized and can be hosted on a university web page.
The Multi-frame Moving Object Detection System enhances remote sensing applications by detecting objects as small as one pixel in low-visibility conditions. It improves signal-to-noise ratio and detects fast- and slow-moving objects with high accuracy.
De-Stijl, a machine learning-based tool, suggests color palettes and demonstrates their impact in different distributions. It removes tedious guesswork from graphic design, facilitating creativity.
Aerovy Mobility, a cloud solutions provider for electric aviation, received $140,000 from the Startup Foundry investment. The company plans to use the funds to launch its live operations tool, VEMS, at vertiports and airports worldwide. Uniform Sierra Aerospace, a drone manufacturer for public safety departments, also benefited from th...
A team of researchers developed a solution using GPT-4 to automate analysis of building operational data, providing comprehensive support for energy management. The codes generated by GPT-4 demonstrate high accuracy in energy load prediction and fault diagnosis.
Lero is recruiting 16 top international post-doctoral researchers for a €2.9 million fellowship program focusing on privacy, trust, inclusion and fairness in software expertise. The program will provide discipline-specific skills training and enhance career development.
Researchers found that many satellites lack basic security concepts, such as separation of code and data, and rely on obscurity to ensure security. The team conducted a thorough analysis of three low-earth orbit satellites and identified vulnerabilities that could be exploited by hackers.
A new machine learning-based simulation method called Materials Learning Algorithms (MALA) has been developed, enabling accurate electronic structure calculations at large scales. MALA achieves this by utilizing a hybrid approach that combines physics-based approaches with machine learning to predict the electronic structure of materials.
A new study found that deepfake video clips prompted false memories of films in 49% of participants, comparable to text descriptions. Many participants reported remembering the fake remakes as better than originals, highlighting concerns over artistic integrity and shared social experiences.
A groundbreaking study using ArchCUT3-D software analyzes ancient rock engravings, uncovering unique techniques and production procedures. The research sheds light on the craftsmanship, artistic expression, and cultural context of ancient cultures.
Researchers summarize existing compiler technologies in deep learning co-design and propose a new framework, the Buddy Compiler, to address performance bottlenecks in current AI applications. The study highlights the importance of hardware-software co-design in achieving optimal efficiency and effectiveness in deep learning systems.
A University of Missouri study found that consumers are more likely to use virtual apparel try-on software if it is interactive and easy to use. The research suggests that making the technology versatile can reduce privacy concerns, with qualities like interactivity having a direct impact on users' willingness to shop virtually.
A new study from the University of Copenhagen and Roskilde University found that computers malfunction between 11 and 20 percent of the time, wasting users' productivity. The researchers identify insufficient performance and lack of user-friendliness as the two biggest categories of problems.
AnalySwift will develop DATC, a design tool for engineers to analyze lightweight structures made from advanced tailorable composites. The tool aims to improve NASA's capabilities in designing and analyzing aerospace structures, reducing the need for costly physical experiments.
Researchers developed AEGIS, the first technique to detect backdoor attacks in robust machine learning models. It improves the trustworthiness of artificial intelligence by analyzing robust models and detecting mixed input distributions for poisoned classes.
Researchers at Chalmers University of Technology have developed open-source software, SuperConga, to explore new superconducting properties and advance quantum computing. The program operates at the mesoscopic level, enabling simulations that can 'pick up' the strange properties of quantum particles.
Researchers have developed a simulation of how people behave on roads based on cognitive theories, accurately reproducing pedestrian and driver behaviors in common scenarios. The model predicts complex underlying cognitive mechanisms that contribute to the difficulty in creating self-driving vehicles.
The PanDrugs2 programme analyzes alterations in tumour genes, relates them to existing available drugs, and reports on the most suitable ones. It handles data on over 4,600 genes and 14,600 chemical compounds, generating information on around 74,000 associations between drugs and genes.
A network meta-analysis of 10 studies found expert panel-based and computer-based aesthetic outcome evaluation scored lower than patient-perceived outcomes. Standardization of tools with racially and culturally inclusive patient-reported measures is needed to improve clinical assessment of breast cancer patients.
Clean technology can help reduce food waste and establish long-term sustainability goals in on-site foodservice providers. Engaging employees in the process is crucial for its success, as it helps measure and quantify waste production.
Researchers propose a four-step process for guiding the use of ChatGPT in education, identifying desired outcomes, determining automation levels, ensuring ethical considerations, and evaluating effectiveness. The study finds that ChatGPT can improve teaching models, assessment systems, and student learning experiences.
Researchers have developed TomoTwin, an AI-based software that accurately identifies and localizes proteins in cells using electron cryo-tomography. This innovation expands the potential of cryo-ET to decipher biomolecule functions and unveil disease origins.
SpaceMarkers, a new machine learning software, can identify molecular interactions among distinct cell types in and around tumors. By analyzing spatial transcriptomics data, it reveals genes overexpressed due to cell-to-cell interactions, providing new avenues for understanding cancer progression and treatment responses.
A team of researchers from University of Missouri and University of Nebraska-Omaha are working on a three-year, $1.6 million grant to make open-source software tools more accessible and understandable to a wider range of people and tech companies. The project aims to identify what OSS projects are good to invest in for commercialization.
Researchers successfully applied reinforcement learning to protein design, creating proteins with improved antibody generation and accurate nano-structures. The approach may lead to more potent vaccines and novel applications in regenerative medicine.
A study at ECCMID shows AI software can detect TB from chest X-rays with sensitivity and specificity comparable to that of trained radiologists. The technology has the potential to improve diagnosis in low-resource settings where radiologists are scarce.
A new study has introduced a technique to effectively and efficiently mitigate risks by misperception of image recognition AI used in automated driving systems. Researchers have developed DNN repair methods that explore sets of parameters responsible for each misperception type, leading to more effective risk mitigation.
OncoMerge uses genetic data to analyze tumor activity and predict future changes. The software detects abnormal gene fusions and mutations affecting protein expression and gene copy numbers, improving the accuracy of cancer modeling predictions.
A new model of DNA flexibility has been developed, providing results of unprecedented quality and characterizing precision and efficiency at the computational level. The study presents a systematic and comprehensive analysis of DNA movement correlations and introduces a new method to capture them.
Researchers at the University of Kansas propose a data project life cycle to equip students with essential data skills, including spreadsheet wherewithal. The approach emphasizes planning, acquiring, organizing, analyzing, and interpreting data in simple spreadsheets.
Researchers at the University of Missouri are acquiring a new transmission electron microscope (TEM) with a $800,000 grant from the National Science Foundation. The TEM will allow them to conduct experiments in real-time and gain a greater understanding of material structure at an atomic level.
Researchers developed an image analysis algorithm that can automatically measure Arabidopsis thaliana stomatal aperture with high accuracy and speed. The technology also includes a portable imaging device for non-destructive observation using intact plants, allowing for rapid measurement of subtle changes in stomatal aperture.
Researchers have developed a high-speed, 3D gigapixel microscope that stitches together dozens of cameras to capture life in unprecedented detail. The device enables the recording of differences in pitch and depth, allowing scientists to study zebrafish behavior and developmental biology without harming the animals.
The CONVINCE project, funded by €4 million, focuses on developing cognitive deliberation capabilities in autonomous robots to enhance safe and robust operation in unstructured environments. The goal is to enable robots to adapt to their surroundings and tasks autonomously without human intervention.