Articles tagged with Artificial Intelligence
The new AIFS ENS model outperforms state-of-the-art physics-based models, with gains of up to 20%, and generates forecasts over 10 times faster while reducing energy consumption by approximately 1,000 times. This complements ECMWF's portfolio by leveraging machine learning and artificial intelligence.
A new study reveals that human-attributed responses are perceived as more supportive and emotionally resonant than identical AI-generated responses. Participants consistently rated 'human' responses as more empathic and satisfying, especially when emphasizing emotional sharing and genuine care.
Researchers at USF have developed a system that can detect distinct patterns in facial movements linked to emotional expression in children with PTSD. The technology uses de-identified data from video analysis to provide an objective, cost-effective tool for clinicians to identify and track PTSD in children and adolescents.
Researchers at Cornell University have developed a process that transforms short videos of rooms into highly accurate, interactable 3D simulations. The technology can be used to create more realistic video games and train robots to operate within specific real-world spaces.
An interdisciplinary team at TU Wien has developed a method that allows for the exact calculation of how reliably a neural network operates within a defined input domain. This enables mathematical guarantees for the safe use of AI in sensitive applications.
A new AI tool, iSeg, has been developed to accurately outline lung tumors on CT scans and identify areas that may be missed by doctors. The study found that iSeg consistently matched expert outlines across hospitals and scan types, and flagged additional areas that some doctors missed.
The AI for Good Global Summit 2025 will showcase AI innovations delivering better healthcare and education, reducing disaster risks, ensuring water and food security, and bolstering economic resilience. The event, organized by the International Telecommunication Union (ITU), features talks from AI leaders and 100+ demos.
The article discusses how AI systems are transforming historical records without transparency around methodology selection, weighting, and interpretation. Historian Marnie Hughes-Warrington argues that AI can be an opportunity to engage with its development to ensure complex histories are reflected.
A new study developed a powerful method for detecting simultaneous nitrogen and water deficiencies in field-grown sesame using deep learning and multiple data sources. The integration of hyperspectral, thermal, and RGB imaging enabled identification of combined nutrient and water-related deficiencies.
A research team at Osaka Metropolitan University has developed an AI model that can detect fatty liver disease from chest X-ray images with an accuracy rate of 0.82-0.83. This breakthrough has the potential to improve early detection and treatment of the disease, which affects one in four people worldwide.
The Arc Institute has launched its inaugural Virtual Cell Challenge, a public competition using AI to solve one of biology's biggest challenges. Competitors will train models on gene expression data from over half a billion cells and predict changes in gene activity when individual genes are silenced.
This issue of SLAS Technology features a high-precision microfluidic flow splitter that outperforms commercial alternatives, enabling even flow division and simplifying multi-inlet perfusion. The journal showcases technological leaps in the life sciences, including rapid pathogen detection and AI-driven insights into schizophrenia.
Academic medical centers can transform patient care by adopting learning health system principles, harnessing real-time data analysis and AI to improve patient outcomes. The proposed model aims to establish a smarter, more efficient and equitable care model by integrating diverse databases and leveraging AI for personalized, proactive ...
Researchers at TU Graz developed methods to run AI models locally on small devices with limited memory, enabling efficient positioning error correction and industrial applications. The E-MINDS project introduced a modular system using division, orchestration, subspace configurable networks, quantisation, and pruning techniques.
Researchers discovered AI art protection tools have critical weaknesses that cannot reliably stop AI models from training on artists' work. LightShed, a new method, can detect and remove distortions, stripping away protections and rendering images usable again for generative AI model training.
Researchers developed an AI-driven framework to improve the mechanical properties of two-dimensional patterned hollow structures (2D-PHS). The framework achieved a 4.3% improvement in stress uniformity and a 23.1% reduction in maximum stress concentrations, increasing tensile strength by up to 12%.
Researchers discovered AI surrealism, cultural bias, logical fallacy and misinformation in image-generated content. Developing standardized language to describe these flaws will help train AI to generate images more accurately.
A new dataset doubles documented stream miles in the Chesapeake Bay Watershed, allowing for more accurate characterization of water flow and land use. The high-resolution data will help prioritize restoration projects, such as streamside tree plantings and pollutant filtering.
The ITU-UNICEF initiative has established a new base in Geneva to connect every school to the Internet by 2030. The centre hosts training and support for governments to meet their digital development goals, promoting inclusive education.
Researchers developed an AI model using FELA and machine learning techniques to assess uplift resistance in cohesive-frictional soils. The model identified embedment depth ratio, load inclination angle, and soil strength ratios as key factors affecting pipeline stability.
The UC Davis-designed payload is a dynamic digital twin that models the current condition and predicts the future condition of the spacecraft's power system. The satellite will monitor its own health in space using sensors to assess voltage and measurements of the batteries it is running on.
A public-private partnership integrates large language models and multimodal AI to automate MBE growth, improving reproducibility and efficiency. The AI software will be tested on Gallium Nitride before being applied to complex materials systems.
Computer scientist Zeynep Akata has developed innovative methods to combine visual, linguistic, and conceptual elements in AI to increase user trust. Her research on explainable AI aims to make image classification decisions more transparent.
A new study reveals a five-fold increase in computer vision papers linked to surveillance patents, highlighting the rise of obfuscating language that normalises surveillance. The top institutions producing surveillance are Microsoft, Carnegie Mellon University, and MIT.
A new Center for Protein Design at the University of Copenhagen aims to create artificially designed proteins with tailored properties to tackle diseases, environmental issues, and industrial applications. The centre will drive fundamental research and translate basic findings into concrete solutions.
Current risk-based regulatory approaches to AI in healthcare fall short, ignoring individual patient preferences and potential for bias. Authors call for introduction of patient rights relating to AI-generated diagnosis or treatment planning to protect patients' autonomy and trust.
The report highlights cutting-edge technologies like generative watermarking, autonomous biochemical sensors, and green nitrogen fixation, which can accelerate progress on urgent global challenges. These innovations represent real, scalable solutions for resilience, health, and equity in the next five years.
The WSIS+20 High-Level Event 2025 in Geneva will assess 20 years of using digital technologies for progress and chart the future direction ahead of the UN General Assembly review. The event will tackle critical issues like the digital divide, AI governance, and sustainability.
Researchers developed an AI tool called AAnet to characterize cancer cell diversity, identifying five distinct cell groups with different gene expression profiles. This could lead to more targeted therapies and improved patient outcomes.
A new study published in PeerJ Computer Science reveals significant accuracy-bias trade-offs in AI text detection tools, which could disproportionately affect non-native English speakers and certain academic disciplines. The research highlights the limitations of detection-focused approaches and urges a shift towards ethical use of LLM...
The Association for Computing Machinery (ACM) has announced a new open-access journal, TAISAP, focusing on AI security and privacy. The journal aims to develop methods for assessing the security of AI models, systems, and environments, including adversarial attacks, privacy concerns, and cybersecurity applications.
The Cirrus Resting State fMRI Software uses AI technology to rapidly map the brain and identify sensitive areas controlling critical functions. The technology, developed at WashU Medicine, enables more precise neurosurgeries with improved accuracy and accessibility for patients.
A new international study found that AI exposure has not caused widespread harm to workers' mental health or job satisfaction, but may be linked to modest improvements in physical health. The research highlights the need for continued monitoring of AI's broader impacts on work and health.
Researchers found that LLMs factor in unrelated information like typos, extra space, and colorful language when making treatment recommendations. This nonclinical information can lead to inaccurate advice for female patients, who are more likely to be erroneously advised not to seek medical care.
BingoCGN accelerates real-time large-scale graph neural network inference through cross-partition message quantization and a novel training algorithm, achieving up to 65-fold speedup and 107-fold increase in energy efficiency compared to state-of-the-art accelerators.
A study found that customized AI chatbots can generate health disinformation, including fake references and scientific jargon. Researchers evaluated five foundational large language models and discovered that four of them consistently provided incorrect responses to health queries.
Cassie, a digital-human assistant developed by Texas A&M University, is transforming the way patients interact with healthcare providers. With facial recognition and emotional intelligence, Cassie offers a two-way interaction that feels like a conversation.
Scientists developed an algorithm that can accurately simulate atomic interactions on material surfaces, reducing the need for massive computing power. This breakthrough enables the analysis of complex chemical processes in just two percent of unique configurations, paving the way for improved battery performance.
Researchers propose a retrieval-augmented generation method based on LLMs and domain KG, achieving surpassing diagnostic capabilities of experienced engineers. The system has been integrated into the CNC Cloud Manager APP, addressing challenges in symbolic reasoning and providing a standardized framework for industrial applications.
Researchers analyzed thousands of viral proteins and found that their bonds to protective antibodies have weakened over time, increasing pandemic potential. The study suggests that candidate vaccines developed 10 years ago may not be efficacious against contemporary strains.
The article proposes a new perspective on AI development, emphasizing the need for consistency in logical structures among datasets, AI models, model-building software, and hardware. The authors suggest integrating the principle of compromise-in-competition into AI design to improve predictive capabilities.
Researchers developed key technologies for precise and high-speed bonding and adhesive technology to address demands of high-performance computing applications. They successfully integrated chips onto a 300 mm waffle wafer, achieving enhanced bonding speed without chip-detachment failures.
Researchers at KAIST have developed a technology to enhance creative generation of AI generative models like Stable Diffusion, generating novel and useful images. The algorithm amplifies internal feature maps to boost creativity without new training, outperforming existing methods in novelty and utility.
Researchers found that some AI prompts create up to 50 times more CO2 emissions than others, with reasoning-enabled models producing the most emissions. Users can significantly reduce emissions by prompting AI to generate concise answers or limiting high-capacity model use.
Researchers developed a technique to study moral decision-making while driving, testing it on 274 philosopher participants. The results showed consistency across different philosophical schools of thought regarding what constitutes moral behavior in the context of driving.
Researchers used nonsense words to test ChatGPT's language processing capabilities, finding it excelled at discovering relationships and providing definitions for extinct words. However, the AI also generated incorrect or made-up answers in some cases, highlighting its limitations.
A new study reveals that large language models exhibit 'position bias', favoring information at the beginning and end of documents or conversations. Researchers identified design choices and training data as contributing factors to this phenomenon, which can be mitigated through adjustments in model architecture and fine-tuning.
A research team developed AI technology that analyzes individual personality traits and values to generate personalized analogies, allowing people to understand others' feelings through familiar experiences. This approach significantly improved emotional understanding and empathy in participants compared to traditional methods.
Researchers at Texas A&M University found that websites use browser fingerprinting to track people across browser sessions and sites. Even users who opt out of tracking under privacy laws may still be silently tracked through fingerprinting.
Researchers have developed a more efficient method for producing green ammonia using artificial intelligence and machine learning. The new process achieves a sevenfold improvement in production rate while being nearly 100% efficient, making it a viable alternative to traditional methods.
A review of AI eye imaging devices approved for patient care found significant gaps in evidence, including lack of transparency on training data and limited diversity in clinical evaluations. The study highlights the importance of rigorous, transparent evidence and data to ensure equitable and effective AI-based solutions.
Researchers developed a new robot navigation system called LENS, which uses brain-inspired computing to set a low-energy benchmark for robotic place recognition. The system combines a spiking neural network with a special camera and low-power chip to enable fast and energy-efficient location tracking.
Researchers at PSI developed an AI-based model to simulate and optimize cement formulations with lower CO₂ emissions. The model, trained on existing data, can generate practical recipe suggestions in seconds, accelerating the development cycle.
Global education leaders call for collaboration, ethics, and human-centered teaching as they confront the benefits and challenges of AI in education. The discussion emphasized the need for responsible use policies, equitable access to AI tools, and preserving uniquely human qualities in education.
Scientists created a low-cost, durable, highly-sensitive robotic 'skin' that can detect various types of touch and pressure. The technology senses multiple physical inputs simultaneously, allowing robots to interact with their environment in a more human-like way.
Researchers at UC Davis developed a brain-computer interface that translates neural activity into speech in real time. The technology allows individuals with ALS to communicate more naturally and inclusively, with 60% of synthesized words intelligible to listeners.
MIT researchers create a novel AI hardware accelerator that performs machine-learning computations at the speed of light, classifying wireless signals in nanoseconds. The photonic chip is scalable, flexible, and energy-efficient, making it suitable for future 6G wireless applications.
Researchers have developed a new method to physically restore original paintings using digitally constructed films that can be removed if desired. The process uses a polymer film mask printed on a very thin film and aligned to an original painting, which takes around 3.5 hours from start to finish.
The UK's first AI for Science Master's programme was launched by King's College London, focusing on interdisciplinary research and education. The programme aims to create a generation of AI-minds ready to tackle scientific challenges, with applications in areas such as biomedicine, chemistry, and materials science.
The Lieber Institute is developing a new tool called GRAPE that combines generative and predictive AI to find new, more effective treatments for brain disorders such as schizophrenia. The institute is utilizing AWS gen AI and compute services to advance research and store its massive collection of genomic and other data in the cloud.