Articles tagged with Artificial Intelligence
The Association for Computing Machinery has named 72 new Fellows for their technical innovations and service to the field. The honorees represent the top 1% of professionals worldwide, with contributions in AI for healthcare, computer graphics, data management, and more.
The Institute of Science and Technology Austria (ISTA) has received a significant donation to advance trustworthy AI technology. The €5 million gift from Garrett Camp will support fundamental research in artificial intelligence, focusing on interdisciplinary collaboration and long-term impact.
The expert consensus provides a structured framework for assessing large language models (LLMs) before deployment in clinical workflows. The framework includes retrospective evaluation, six key capability domains, and essential safeguards for patient data protection and bias mitigation.
The journal has been recognized for its significant international citations, with over 3 million documented in the past year. Its editorial leadership and publishing model ensure rigorous peer review and a multidisciplinary perspective, supporting high-quality research on AI and clinical applications.
A new special issue explores how artificial intelligence can enhance soil carbon storage and ecosystem health in farmlands, forests, and grasslands. The journal Biochar publishes original studies on biochar production, processing, and applications.
A new study by Penn State researchers found that the level of transparency in artificial intelligence (AI) systems affects users' trust. When systems meet or exceed users' expectations, trust increases without needing detailed explanations. However, when expectations are not met, users require more information and a deeper understandin...
A new study reveals that around one-third of all newly written software functions in the United States are already being created with AI support, with productivity gains mainly driven by experienced developers. The study found wide regional gaps, with the US leading at 29% AI-assisted code, while China and Russia lag behind.
The event highlighted urgent global risks and science-led solutions for antimicrobial resistance, trust in data, and AI-enabled sustainable cities. Experts emphasized the need for immediate actions against AMR, which could kill more people than cancer by 2050.
The Center for AI Responsibility and Research will build upon Binghamton research advancing AI for the public good. The new center is supported by a $30 million commitment from Tom Secunda and $25 million from Gov. Hochul and the NY State Legislature.
Researchers at Tohoku University's WPI-AIMR propose Harvested Reservoir Computing framework that harnesses road traffic dynamics for energy-efficient AI systems. The approach yields better prediction accuracy before congestion begins, reducing computational overhead and energy consumption.
Researchers used CT scans and AI to find links between poorer abdominal muscle measures and higher incidence of falls in middle-aged adults. Maintaining good core strength throughout adulthood may provide benefits that start in midlife and continue into older adulthood.
Researchers created a new frequency-aware approach to crafting adversarial images that better match human visual perception. The method, called Input-Frequency Adaptive Adversarial Perturbation (IFAP), significantly outperformed existing techniques in structural and textural similarity.
The new platform, led by PPPL, aims to speed up simulations needed to advance fusion energy research. STELLAR-AI will integrate CPUs, GPUs, and QPUs to tackle the challenges of private fusion companies, enabling faster design and optimization of stellarator devices.
LeapSpace is a research-grade AI-assisted workspace built on trusted scientific content, providing transparent and explainable results. It helps researchers uncover new insights, save time, and improve collaboration, while supporting critical thinking and rigorous transparency.
A large-scale study reveals that generative AI models have reached the threshold of average human creativity, but the most creative individuals still outperform even the best AI systems. The study also highlights the importance of human guidance and parameterization in modulating AI creativity.
The new AI, Health, and Digital Spaces section features an editor’s note emphasizing interdisciplinary collaboration between public health experts and AI researchers. It provides a dedicated scholarly home for work integrating domain knowledge with technical innovation.
The Global Brain Economy Initiative aims to establish brain capital as an essential asset for the 21st century, connecting neuroscience with economic policy. The initiative's mission is to address disparities in support for brain capital across various sectors and promote long-term growth, workforce resilience, and social well-being.
A recent survey study found a significant association between generative AI use and increased depressive symptoms in US adults. The magnitude of the effect varied by age group, suggesting that further research is needed to understand the causal relationship and underlying factors.
Researchers at the University of Oxford have created magneto-sensitive fluorescent proteins that can interact with magnetic fields and radio waves. The breakthrough uses quantum mechanical interactions within proteins to enable practical technologies.
Researchers are working on a £3.5 million UK-Japan project to develop human-centred artificial intelligence for sign language translation, focusing on natural conversational Deaf data rather than interpreted signing.
A new medical large language model (LLM) achieved over 91% accuracy in identifying female participants diagnosed with major depressive disorder after analyzing short WhatsApp audio recordings. The LLMs showed greater accuracy when classifying women compared to men, particularly when given the 'describe your week' data.
Protein Foundation Models (pFMs) leverage vast amounts of sequence and structural data to predict protein structures and functions, enabling novel protein design and analysis. The models have evolved into several mature technical approaches, demonstrating versatility in basic biological research, protein discovery, and biomedical appli...
Researchers found that complexity, including ground truth, real-world complexity, and stakeholder involvement, are key factors in reducing AI biases. By accounting for these complexities, developers can create more fair AI models.
A Stanford Medicine-led study develops an AI tool that predicts prematurity complications by analyzing blood samples collected at birth. The algorithm correlates molecular patterns with health outcomes, offering a new understanding of prematurity as multiple distinct conditions.
The AI Polis platform offers a city-level coordination system to design, prioritize, and scale projects that improve urban life while staying within planetary boundaries. By connecting technology, policy, finance, and environmental science, it guides cities in strengthening communities while protecting the planet.
A study found that women are more averse to risk in general and personally exposed to AI risks, but support AI use nearly as strongly as men when benefits are guaranteed. Women's skepticism towards AI is linked to higher risk orientation and exposure to AI-related jobs.
BioPathNet is an AI method that analyzes large biomedical knowledge graphs to identify hidden connections between genes, diseases, and drugs. The model proposes hypotheses that can be tested experimentally or clinically, providing a hypothesis-generating tool for researchers.
Researchers at Sandia National Laboratories have successfully employed artificial intelligence labmates to improve the control of LED lights, leading to a fourfold increase in steering efficiency. The AI platform uses a combination of machine learning and equation-learning techniques to optimize experiments and achieve new insights int...
A new AI model, ShapKAN, has been integrated into the cloud-based platform AI4Min-PE to predict five critical thermodynamic parameters across extreme pressures. The platform offers speed and accuracy far beyond conventional approaches, enabling scientists to explore the chemistry of materials under extreme conditions.
Researchers developed an autonomous AI network technology that enables multiple AI systems to spontaneously collaborate and share knowledge to discover new materials. This collaboration improves exploration efficiency and paves the way for the creation of numerous new materials.
A new AI-based learning tool, NeuroBridge, aims to bridge the communication gap between neurotypical and autistic individuals. The interactive LLM tool evaluates user responses to match autistic communication preferences, teaching users about cross-neurotype communication styles and their interdependent nature.
Researchers created a computational model that combines physiological signals, sensory input, and word information to construct human emotions. The model achieved an agreement rate of about 75% when compared to participants' self-reported emotional evaluations.
Researchers developed a wearable Revoice device that decodes speech signals and emotional cues to enable seamless real-time communication in people with post-stroke dysarthria. The device achieved a word error rate of 4.2% and sentence error rate of 2.9%, showing promising results for stroke rehabilitation.
The ADVANCE-6G, TEST-6G, and INES projects have provided IMDEA Networks with robust experimental platforms to tackle the technological challenges of future networks. The Institute has strengthened its research capabilities in advanced 5G and future 6G networks.
A team of researchers used Dungeons & Dragons as a testing ground for Large Language Models, evaluating their performance in long-term tasks. The models showed quirky behaviors but excelled in staying 'in character' and determining correct actions, with Claude 3.5 Haiku performing the best.
A new study reveals AI is being used to recreate deceased people's voices, faces, and personalities without consent, raising questions about ownership, exploitation, and power. The study identifies three dominant ways AI brings back the dead, including spectral labor, where digital remains are extracted and monetized.
Researchers developed AI-powered in silico labeling to analyze cell images without staining, preserving cell health. The system leverages context, such as cell shape and position, to accurately stain rare processes like cell division.
Researchers developed an energy learning hyper-heuristic algorithm to tackle complex task assignment challenges in heterogeneous UAVs. The algorithm outperforms traditional methods in convergence speed and solution quality, enabling efficient completion of tasks while avoiding obstacles.
Researchers have discovered the ExHOS nanomachine, which controls constitutive exocytosis by delivering spherical molecular packages to the cell surface. This process is essential for preserving cell fitness and vital functions such as communication and growth.
The University of Cincinnati College of Medicine has been awarded a $1.1 million AMA Transforming Lifelong Learning Through Precision Education grant to develop an AI-powered system for personalized learning and feedback. The project aims to improve the quality of medical education by providing real-time feedback on clinical reasoning ...
Developed by Mass General Brigham researchers, the autonomous AI system achieves high accuracy in detecting cognitive impairment using routine clinical documentation. The system includes five specialized agents that critique each other and refine their reasoning, just like clinicians would in a case conference.
A new study predicts train-induced environmental vibration using physics-based numerical simulations and limited field data, outperforming conventional models. The method improves existing workflows, reducing dependence on expensive field measurements.
Researchers at IIT develop optical microscopy technique that combines polarization and dark-field microscopy to observe cells with high contrast, preserving their natural conditions. The next step involves using AI to enrich images with molecular information related to diseases.
A chemist proposes a framework for shared model proteins to improve reproducibility and coordination in protein science. The proposal includes five widely used proteins and aims to establish minimal reporting requirements and curated reference datasets.
The article reviews AI's potential to improve mental health diagnoses, monitoring, and treatment while addressing unique challenges such as biased training data and patient privacy concerns. The authors argue for human-centered approaches and coordinated oversight to ensure responsible AI deployment in mental health care.
Engineers at the University of Pennsylvania have discovered that foams exhibit internal motion resembling deep learning in AI systems. The study suggests a common mathematical principle underlying both foams and AI training, with implications for designing adaptive materials and understanding biological structures.
Researchers developed a new technique called CLASSIC that enables large-scale testing of complex DNA circuits in human cells. The approach uses artificial intelligence and machine learning to analyze vast numbers of complete circuits at once, providing scientists with a clearer picture of the rules governing genetic part behavior.
AI tools enable scientists to publish more papers, receive citations, and become leaders earlier, but also narrow the scope of science by concentrating attention on data-rich domains. This 'lonely crowds' phenomenon reduces interaction among researchers, leading to overlapping research and a contraction in knowledge extent.
Scientists created biologically realistic artificial cilia using hydrogel, enabling precise control over their motion. The tiny structures can be powered by low-voltage electrical signals and have shown remarkable durability and versatility.
Oliver Zielinski's contributions to oceanography have significantly advanced understanding and stewardship of the ocean through innovative research and leadership. He has led applications of artificial intelligence for environmental monitoring and mentored early-career scientists.
Researchers developed EnCompass, a framework that enables programmers to improve agents' performance by allowing easy recovery from errors and testing different search strategies without major code rewriting. This improves the organization of core logic in complex AI systems.
Researchers have decoded the logic of microRNA strand selection using AI, revealing a conserved and programmable mechanism governing gene regulation. The study found that this decision follows conserved rules rather than chance, with mammalian microRNAs showing a strong bias towards a single dominant strand.
A team of researchers from Keio University has developed a novel system that uses Gaussian process regression to model and reproduce complex human motions. This enables robots to adapt to touch and interact with objects in dynamic environments, improving their dexterity and versatility.
Researchers develop a framework that mirrors human expert assessment to verify AI-generated answers, filtering out clear mistakes and identifying nuanced gaps. The system saves thousands of working hours and highlights the importance of contextual knowledge and internal documentation.
Large language models (LLMs) provide treatment recommendations for early-stage hepatocellular carcinoma (HCC) that align with clinical guidelines. However, their performance is limited in late-stage disease, where they prioritize tumor factors over liver function. LLMs should be used as a supplement to clinical expertise.
Researchers developed ZENN, an AI framework that recognizes and adapts to hidden differences in data quality. By using thermodynamic principles, the model distinguishes meaningful signals from uncertainty and noise.
Researchers developed PDJA, a novel AI-based attack framework that amplifies vulnerability and bypasses defenses in multi-agent reinforcement learning systems. The approach improves attack effectiveness and cross-layer vulnerability exploitation, opening new opportunities for evaluating the robustness of AI-driven autonomous systems.
A study published in PNAS found that over 60% of pairs of damaging genetic variants can restore enzyme activity when combined, challenging long-held assumptions about genetics. This phenomenon, known as intragenic complementation, has implications for rare disease diagnosis and treatment.
Researchers at Colorado State University used AI to modify antibodies into stable intrabodies that can visualize histone modifications in real-time. This allows for better understanding of gene expression and its relationship with cancer and other disorders. The team created 19 new antibody-based probes with a 70% success rate, signifi...
A machine-learning-based algorithm developed by Tokyo Metropolitan University researchers can accurately count sister chromatid exchanges (SCEs) in chromosomes, giving a more objective measurement. The accuracy rate is 84%, which could help diagnose disorders like Bloom syndrome with greater consistency.