Articles tagged with Artificial Intelligence
Researchers developed a microfluidic platform that squeezes individual breast epithelial cells to measure their mechanical age, revealing an unexpected insight: older cells are stiffer and at higher risk of cancer. The AI-powered platform provides a non-genetic test for women with unknown genetic risks.
Recent AI tools offer surgeons assistance in complex decision-making by analyzing donor hearts and providing a data-driven approach. This could lead to increased efficiency in the donor process, reducing the likelihood of hearts going unused due to time constraints.
Researchers found that firms adopting GitHub Copilot increase hiring of software engineers, with new hires exhibiting more non-programming skills without sacrificing coding ability. The study suggests generative AI is expanding opportunities in the software workforce, not narrowing it.
A recent study by AI researcher Oskar van der Wal found that language models like ChatGPT can absorb biases around gender and ethnicity, which are then embedded in the model. To detect and remove these biases, new measurements are needed.
Researchers found that modern AI language models can distinguish between categories like commonplace, improbable and impossible events with high accuracy. The models' internal patterns, or vectors, show a correlation with human uncertainty about statement plausibility.
A Tohoku University research group developed an AI model that estimates retinal age from a fundus photograph, reflecting biological aging and potential disease risk. The model is non-invasive and can be used as a screening aid, identifying patients who may need further health assessments.
Researchers developed a Variational Level Set Autoencoder (VLSet-AE) to automate contour recognition in SEM cross-sections of DRIE structures. The model achieved high precision, recognizing critical structural features with low average prediction error of 3.65% and correlation coefficient of 0.998.
The University of Manchester has been recognized for its significant contribution to computer science with the third IEEE Milestone Award, honoring the invention of Manchester Code in 1948-1949. The code's self-clocking design enables reliable transmission and remains a key feature in modern digital systems.
Researchers at Saarland University are developing smart implants that can continuously monitor and visualize the healing process of fractures. These customized implants can dynamically adapt to the healing process by becoming stiffer or more compliant as required, promoting bone regeneration through micromechanical stimulation.
Terrence Sejnowski receives Scientific Breakthrough Award for his foundational development of Boltzmann machines, providing the architectural bedrock for deep learning and generative AI. His work has had a profound impact on modern artificial intelligence and tools like ChatGPT.
Binghamton University's new initiative aims to advance artificial intelligence for the public good by educating students on foundational AI principles, workforce applications, and ethical considerations. The three-year, $900,000 program includes a free online microcredential and student research opportunities.
A new study reveals that AI's strength lies in teamwork with humans, processing data but relying on people for context and ethical decisions. The research emphasizes the importance of human oversight and responsibility in AI-driven decision-making.
A new method called scSurvival uses single-cell genetic data to identify which cells inside a tumor are most strongly linked to patient survival. The approach pinpoints harmful and helpful cell populations that can drive disease progression, enabling better understanding of why patients with the same cancer have different outcomes.
Researchers developed a cancer assessment tool that can identify high-risk patients and specific cell populations linked to their risk. The tool, called scSurvival, predicts survival outcomes more accurately than traditional methods by analyzing single-cell data at cellular resolution.
The research team developed a novel AI pathology analysis system named PRET that can accurately recognize multiple types of cancer using only a minimal number of samples. PRET outperformed existing methods in 20 tasks, achieving high diagnostic accuracy rates and stable generalizability across different populations and regions.
The new principles aim to guide boards in navigating the opportunities and risks of AI, with a focus on accountability, trust, and long-term value creation. They cover five key areas: strategic oversight, active technology and security oversight, workforce transformation, building trustworthy AI, and the board's role in governance.
A recent study from Lancaster University reveals that AI systems like ChatGPT can learn to mirror human impoliteness, potentially escalating into verbal violence. The research tested ChatGPT's ability to respond to real-life impolite interactions, finding it often produces more impolite behavior than humans.
A recent survey found that 72% of youth aged 13-17 use AI companions and 52% report regular use, highlighting the need for robust safeguards. Well-designed chatbots can normalize help-seeking, reduce isolation, and offer coping strategies, but poorly designed ones can cause harm.
The article discusses how emerging digital tools are capturing the biopsychosocial reality of chronic pain. Digital tools such as wearables, AI-driven trackers, and ecological momentary assessments mitigate recall bias by recording data in real-time, providing a more holistic picture of the patient's journey.
Researchers review advanced AI algorithms and hardware acceleration techniques to predict material properties, optimize structures, and discover new materials. This review provides crucial guidance for accelerating data-driven materials research and fostering next-generation functional materials development.
Artificial synapses are built from soft, bio-friendly materials that operate like human brain synapses, merging data storage and computing into a single unit. Laboratory prototypes demonstrate immense capabilities, consuming energy on the scale of femtojoules.
A new coaching tool helps users identify biases in their prompts and generate more inclusive content. The study found that the intervention increased users' awareness of algorithmic bias and boosted confidence, but led to a less satisfactory user experience.
A recent study published by the American Psychological Association found that people who rely heavily on AI programs for work tasks experience reduced confidence in their own independent reasoning. In contrast, those who actively challenge or modify AI suggestions report greater confidence and a stronger sense of authorship.
Indiana University has expanded its free generative AI course, GenAI 101, to over 805,000 alumni worldwide. The course provides practical skills in prompt engineering, data storytelling, and ethical AI use, preparing learners for an AI-powered world.
Researchers at Tohoku University developed an AI-based method integrating physics-based modeling for rapid screening of material candidates. The approach significantly improves accuracy by evaluating basic properties before predicting complex ones.
Dr. André Biedenkapp's work on generalizability in reinforcement learning aims to make AIs more robust and adaptable, with potential applications in real-world scenarios
Artificial intelligence models provide personalized advice, but may perpetuate negative stereotypes about people with autism. Researchers found that up to 70% of the time, AI discourages those with autism from socializing.
Emerging AI technologies are enabling faster, smarter, and more integrated solutions to global challenges. AI-powered systems can track pollution levels, detect anomalies, and predict future risks in water, soil, air, and waste systems.
The University of Chicago's Data Science Institute is developing AI-based forecasting technology to support farmers and citizens worldwide. The project aims to deliver forecasts that inform decision-making on agriculture, public health, and extreme heat avoidance.
Engineers at Northwestern University developed artificial neurons that generate realistic electrical signals to activate living brain cells. This breakthrough paves the way for brain-machine interfaces and neuroprosthetics, as well as more efficient brain-like computing systems.
A new study found that AI applications enhance consumer understanding of skin conditions by increasing accuracy and confidence. However, imperfect guessing accuracy remains, highlighting the need for further design improvements to facilitate better comprehension.
Researchers developed a new AI-assisted tool called CacheMind to improve cache performance and reduce evictions. The tool uses causal reasoning to analyze fine-grained details about system behavior, enabling computer architects to identify patterns and implement fixes.
A team of scientists from NTU Singapore has developed a new biochip that, when paired with Artificial Intelligence (AI), can detect quickly and accurately extremely small amounts of microRNAs. The device can cut detection time from hours to 20 minutes.
Recent progress in advanced energy manufacturing highlights 3D printing's potential to redefine next-generation lithium batteries. The technology enables precise control over three-dimensional structures, improving ion-transport pathways and mechanical robustness.
The researchers developed an AI-based method that allows users to input natural language prompts about the materials they want to create and suggests optimal procedures for experiments to produce them. The method has been successfully applied to identify catalysts for turning carbon dioxide and hydrogen into carbon monoxide and water u...
Researchers propose a novel brain architecture for efficient processing, integrating parallel cortical and subcortical pathways. This approach may improve decision-making tasks, suggesting current AI models are missing key brain function principles.
Researchers identified a practical strategy to reduce bias in AI systems used in children's mental health care. By analyzing training data, they found that targeted improvements can significantly reduce disparities without diminishing overall model performance.
Advances in high-throughput proteomics and artificial intelligence are transforming biomarker discovery, disease prediction, and drug development. Proteomic technologies enable comprehensive snapshots of physiological and pathological states, driving breakthroughs in early diagnosis and personalized treatments.
The Journal of Management Analytics seeks high-quality research on the intersection of artificial intelligence and business disciplines, focusing on prescriptive intelligence to optimize organizational systems. The symposium will be held in June 2026 at ESCP Turin, Italy, with partial funding available for selected authors.
Ricardo Baeza-Yates has made fundamental contributions to computer science through his pioneering work in algorithms and information retrieval. He has fostered a vibrant transnational research community across Latin America, inspiring young people with the South American Programming Contest.
A new computational model developed by researchers at Chalmers University of Technology shows that onshore wind is likely to supply around 25% of global electricity by 2050 and solar reaching about 20%, aligning with the 2°C target. However, this falls short of what is required for a 1.5°C goal.
The study reveals that metabolic dysfunction-associated steatotic liver disease (MASLD) is emerging as a major contributor to liver cancer worldwide and in China. Researchers call for integrated global prevention and early detection strategies to address this growing burden.
Researchers from Juntendo University found that AI-based assessment of medical interview transcripts is comparable to clinicians' scores while reducing evaluation time. The study demonstrated strong agreement between AI and human evaluations, with AI showing greater consistency across repeated assessments. This suggests a potential 'AI...
A study by Hector Zenil and colleagues reveals that perfect AI alignment with human values is mathematically impossible. Instead, they propose 'managed misalignment' as an alternative strategy, utilizing competing AI agents with different cognitive styles to dynamically aid or thwart one another.
Researchers used 3D imaging and artificial intelligence to analyze the microscopic structure of coral skeletons, revealing subtle changes caused by Stony Coral Tissue Loss Disease. The study found that Attention U-Net performed best in detecting differences between healthy and diseased corals.
A research team has developed a comprehensive dynamic Gaussian processing framework that provides a complete pipeline from data acquisition to practical application. The framework addresses the high storage cost and fragmented tool chain ecosystem associated with volumetric video technology.
The University of Chicago Data Science Institute has welcomed Millennium as a new Industry Affiliate Partner, aiming to advance curriculum and drive innovation. This partnership will support the design of a new quantitative developer certificate program for students.
The University of Chicago has partnered with AI Research Commons and Microsoft to support high-potential AI startups emerging from the Third Coast Foundry universities. Selected companies will receive technical and go-to-market support, up to $350K in startup credits, and connections to the Bay Area AI investor network.
Global warming poses a significant threat to rice production, which has historically been heat-loving but is now facing its limits. Rice can tolerate temperatures below 82°F and 104°F, but rising temperatures will accelerate beyond these thresholds.
Researchers at the University of Washington developed a system that integrates tiny cameras in wireless earbuds with an AI model, allowing users to ask questions about the scene in front of them. The system achieved high accuracy rates for translation and object recognition, making it a promising technology for future applications.
A machine learning technique identifies optimal spots for tree planting to inform climate mitigation strategies. Optimised afforestation can reduce river peaks by up to 43% while preserving groundwater by up to 60%. The study's findings highlight the importance of data-driven technologies in supporting large-scale, smarter planning.
A team of Weill Cornell Medicine investigators is creating a dual-track program to train clinical oncology fellows and cancer research scientists in AI tools and cancer biology. The goal is to develop scientists who can use the latest tools to deliver on the promise of personalized cancer therapies.
A study of 21 large language models found that they excel at accurate final diagnoses but falter at differential diagnoses, highlighting the need for human oversight in healthcare. The researchers developed a novel measure, PrIME-LLM, to evaluate AI's clinical competency across different stages of reasoning.
Current large language models struggle with early diagnostic reasoning and cannot be relied upon for patient-facing clinical decisions. The study highlights the need for further improvement in these areas to enhance patient care.
The new guidance provides practical routes forward for AI use in every publishing role, promoting accountability, transparency, risk-awareness, and innovation. It introduces structured checkpoints and ready-to-use tools to make responsible AI use practical, not theoretical.
A Drexel University study found that teens are concerned about their attachment to AI chatbots, using them for emotional support and entertainment. The study suggests that these interactions can evolve into addiction-like behavior, affecting their lives offline.
New research suggests that AI-powered price discrimination could lead to hidden, personalised pricing that consumers cannot see or understand, potentially reducing trust and affecting behavior. Regulators are now facing pressure to decide whether such practices cross the line under EU competition law.
Researchers developed a new training technique, HarmonyGNN, to improve the accuracy of graph neural networks in heterophilic graphs. The framework achieved state-of-the-art performance on four heterophilic graphs with accuracy improvements ranging from 1.27% to 9.6%.
A new study analyzed thousands of textual descriptions of two-person social interactions using generative AI, resulting in a comprehensive taxonomy of categories. The findings provide a data-driven framework for understanding the structure of social situations, including associations with conflict, power, and duty.
The study, led by Lucy Liu and L. Mahadevan, shows that adding the right amount of noise to individual robot movements can ease gridlock and improve efficiency in crowded environments. The researchers used computer simulations and experiments with small robots to test their ideas.