Articles tagged with Artificial Intelligence
A study published by Michigan Medicine suggests that AI can help increase the chance of people with risky drinking patterns or signs of alcohol use disorder getting outreach and help. The NLP tool analyzed full text of health records and identified over 47,500 patients with some sign of risky drinking.
Researchers trained AI systems on data from two cultural groups playing a video game, finding that the agents learned and applied the group's degree of altruism. This approach, called inverse reinforcement learning, more closely aligns with human development and learning, where values are 'caught' rather than 'taught'.
A new study by researchers at TU Graz and the Know Center shows that AI-powered recommender systems can perpetuate unfairness if not designed with human decision-making in mind. The study proposes a multi-stakeholder approach, involving service providers, municipalities, and local residents, to ensure fairness and transparency.
A new study suggests that a simple ECG reading from a smartwatch can accurately estimate a person's age, potentially offering a secure alternative to facial recognition or ID checks. The method achieved an average error of just under three years and accuracy consistently reached between 93 and 96%.
MIT researchers developed a new fabrication method to stack multiple functional components on top of one existing circuit, reducing energy wasted during computation. The new approach enables the production of more energy-efficient electronics, boosting computation speed and reducing electricity consumption.
A new method has been developed to generate valid confidence intervals for problems involving data that vary across space, helping researchers trust the results of certain experiments. This work can be applied to fields like environmental science and epidemiology.
A third of teenagers use AI companions for social interaction, with some reporting more satisfying conversations than humans. Experts caution against problematic chatbot use as a new environmental risk factor, particularly among vulnerable populations.
Scientists at Michigan Medicine have developed machine learning models that can analyze blood samples for biomarkers to detect ALS earlier and predict disease severity. The models show promising results, with predictions accurate up to 91% in some cases, and could lead to improved diagnostic accuracy and treatment options.
The SWIFTT project presents a hybrid seminar to showcase its platform that helps foresters detect and prevent insect outbreaks, analyse windthrow and fire damage. The event features live demos and presentations from the project team, highlighting the platform's benefits for sustainable forest management.
Researchers at Florida Atlantic University have developed a deep learning model that detects and evaluates Alzheimer's disease (AD) and frontotemporal dementia (FTD) using EEG brainwave analysis. The model achieved over 90% accuracy in distinguishing individuals with dementia from cognitively normal participants.
Researchers develop an integrated pipeline to analyze and visualize NHANES dental caries data, identifying age-driven heterogeneity and novel associations between caries status, environmental factors, and disease markers. The study demonstrates the value of integrating machine learning with large-scale observational data for deeper ins...
The startup aims to bring human integration and industrial usability to its robots, which will operate safely and efficiently in industrial environments. The company's mission is to create a future where humans and robots collaborate daily.
Overleaf is introducing a new AI assistant that combines advanced capabilities with its intuitive editor. The tool helps researchers spend less time coding and more time producing high-quality work. The new features are available in Overleaf Labs and provide a research-focused writing assistant and a first-of-its-kind citation reviewer.
Researchers in Australia and Aotearoa New Zealand will have expanded open access publishing options through a new three-year agreement. The partnership supports fully open access journals and enables the expansion of AI use in teaching and research.
Researchers developed multimodal models integrating imaging, clinical, and molecular data from TAILORx tissue biorepository for early-stage breast cancer. The models demonstrated enhanced prognostic performance compared to existing methods, highlighting their potential for personalized treatment decision-making.
The company's AI tool, RECTIFIER, uses generative AI to screen patients for clinical trial eligibility, significantly outperforming manual screening in accuracy and efficiency. The tool has been shown to more accurately identify eligible patients and reduce the burden of manual chart review in various healthcare settings.
Researchers developed an AI tool to identify patients with undiagnosed Alzheimer's disease using electronic health records, addressing underdiagnosis and healthcare inequities. The model achieved sensitivity rates of 77-81% across diverse populations, promoting fairness while maintaining high accuracy.
A new report highlights the surge in child sexual exploitation and abuse online, with AI-generated material increasing by 1,325% in a year. The WeProtect Global Alliance's Global Threat Assessment 2025 provides a practical tool for technology companies and governments to tackle the crisis.
A new multi-modal AI framework, DriveMLM, integrates perception, planning, and human instruction to produce aligned behavioral planning states for autonomous driving. It outperforms existing systems in closed-loop evaluation, demonstrating strong open-loop decision accuracy and robust performance under natural-language guidance.
University of Utah engineers developed an autonomous approach using proximity and pressure sensors, training an artificial neural network on grasping postures. Study participants demonstrated improved grip security, precision, and less mental effort when using the AI-enhanced prosthesis.
The University of Texas at Dallas has partnered with Tech Mahindra to facilitate collaboration on artificial intelligence (AI) innovation, skill development, and research. The partnership will provide opportunities for students and faculty to advance AI technologies, data science, and cybersecurity.
Researchers at the University of Washington have developed smart headphones that use AI to detect conversation cadence and isolate participants in a noisy soundscape. The prototype was tested with 11 participants, who rated filtered audio more than twice as favorably as baseline sound.
The UAE has launched an AI ecosystem to support climate-resilient farming in vulnerable communities worldwide. The CGIAR AI Hub and four other initiatives aim to drive digital transformation and innovation in agriculture, providing practical solutions for farmers, governments, and development actors.
Researchers have developed an AI-driven method to uncover genetic interactions driving cancer progression. The approach highlights complex interplay between genes allowing malignant cells to gain momentum and reveals previously hidden cancer drivers.
Researchers at AIMR and UC Santa Barbara develop a breakthrough digital p-bit design that eliminates bulky analog components, enabling self-organizing hardware-based probabilistic computing. This advances applications in AI, logistics, scientific discovery, and future computing systems.
Researchers at Cold Spring Harbor Laboratory developed an AI-powered approach to identify redundant genes in plants. By analyzing evolutionary data and machine learning models, they predicted which genes to edit to modify specific traits, providing a new 'roadmap' for plant breeders.
Khalessi will shape UC San Diego Health's innovation strategy and lead the integration of new technologies, including AI. He aims to accelerate innovations supporting clinicians, patient safety, and operational efficiency.
A recent study published in Nature Communications found that the human brain processes spoken language in a sequence mirroring advanced AI language models. The research suggests that artificial intelligence can offer insights into how the brain constructs meaning, challenging traditional rule-based theories.
A new model developed by Osaka Metropolitan University Assistant Professor Takuya Fujinaga enables robots to accurately pick tomatoes, with an 81% success rate. The system evaluates the ease of harvesting for each tomato, taking into account factors such as fruit clustering and stem geometry.
Researchers at Stanford University have successfully used machine-learning-based control to operate a robot on the International Space Station, allowing for faster and more efficient navigation while maintaining safety. The system, which includes a warm start feature, enables robots like Astrobee to move through complex environments wi...
A new machine learning model developed at the University of Michigan uses data from wearable sensors to predict how physical therapists would assess balance training performance. The model was trained with sensor data combined with analysis from physical therapists and predicted patients' balance ratings with nearly 90% accuracy.
A new AI-based method can differentiate between tumour progression and radiation necrosis with high accuracy, enabling more accurate identification and treatment of late-stage cancer patients. This discovery could lead to better treatments for brain metastasis patients treated with stereotactic radiosurgery.
A University of Warwick-led analysis of 4,820 student reports found that student writing has become more sophisticated, formal, and positive since the introduction of ChatGPT in 2022. However, grades have remained stable despite stylistic shifts, suggesting core academic skills remain essential.
A study finds that most people trust doctors more than AI for health diagnoses, but have a positive view of AI's potential in cancer diagnosis. The research used two nationally representative surveys to assess public attitudes toward AI in healthcare, including trust, understanding, and excitement.
A breakthrough AI system called OmniPredict can predict human pedestrian behaviors with unprecedented accuracy, revolutionizing self-driving cars and urban mobility. The model combines visual cues with contextual information to anticipate pedestrians' next moves, reducing the risk of accidents and improving traffic safety.
The University of Phoenix College of Doctoral Studies has released a new white paper exploring the leadership opportunities for AI to increase employee value. The authors suggest that leaders can redesign roles, address skills gaps, and unlock higher-value work with generative AI.
The Battery Large Model system revolutionizes battery design, manufacturing, operation, and recycling through AI-simulation synergy. It provides a novel technological path for the industry's intelligent upgrade, enabling autonomous design scheme generation, accurate performance prediction, and intelligent defect detection.
Researchers propose AI-based surrogate brain framework that integrates forward modeling, inverse problem solving, and evaluation to build next-generation data-driven models. The framework enables applications such as predicting future whole-brain activity, performing dynamical systems analysis, and guiding neurostimulation strategies.
A new study presented at the 2025 American Society of Hematology meeting revealed that subtle disruptions in genome architecture can predispose individuals to lymphoma. The research found that damaged 3D genome architecture can silence key tumor suppressor genes, creating fertile ground for malignancy.
The SECURED project aims to provide a secure environment for decentralized, cooperative processing of health data and anonymisation assessment. The consortium is developing AI models to diagnose and treat patients with blood malignancies in a secure way, ensuring patient data protection and compliance with European privacy regulations.
A new study analyzes 183 scientific publications to highlight gaps in AI research for STEM education. The analysis reveals a technocentric approach that neglects non-cognitive skills, ethical issues, and geographical diversity.
A new study from Michigan Medicine found that its data-driven command center increased hospital bed use efficiency by 63%, reducing emergency department boarding time by 33% and patient transfers by 13%. The initiative also streamlined discharge processes, resulting in a 12% drop in waiting times for patients ready to leave the hospital.
A new AI model can accurately identify brain abnormalities in MRI scans for various conditions, including stroke, multiple sclerosis, and brain tumors. The model can also triage scans, increase reporting speeds, and support clinical decision-making by suggesting findings to radiologists.
A new UNC-Chapel Hill study uses large language models to accelerate georeferencing in plant specimen collections, unlocking millions of records for global ecological research. The breakthrough technology improves upon traditional methods with near-human accuracy and efficiency.
IMDEA Networks has created a secure watermarking tool called FreqyWM, allowing institutions to tag their data with a unique signature and preventing leaks and unauthorized copies. The tool enables the exchange and reuse of information while complying with legal frameworks on security and privacy.
The symposium will feature experts sharing cutting-edge research and insights on AI in clinical settings, aiming to improve chronic health outcomes and expand access to mental healthcare globally. The event is a unique platform for networking and collaboration, highlighting innovative approaches that move the field forward.
New research from Cornell University finds that AI chatbots powered by large language models can effectively persuade voters in either direction, shifting opinions by significant margins. The models' persuasiveness comes from providing many factual claims supporting their arguments, but these claims are not necessarily accurate.
Researchers at MD Anderson have made significant discoveries in the treatment of rare bile duct cancers, with zanidatamab showing promising results. Additionally, a study identified RASH3D19 as a target to overcome treatment resistance in KRAS-mutant cancers.
A11yShape, an AI-assisted tool, enables visually impaired computer programmers to create, edit and verify complex 3D models without assistance. The tool uses GPT-4o to provide detailed descriptions of the model for blind and low-vision programmers.
Researchers developed instance-adaptive scaling framework that uses process reward model to estimate difficulty of question, enabling LLMs to spend more compute on promising solution paths. This approach achieves comparable accuracy with existing methods while reducing computational cost by up to one-half.
The MIT team developed a new AI-based controller that enables the robot to follow gymnastic flight paths, such as executing continuous body flips. The robot's speed and acceleration increased by 450% and 250%, respectively, compared to previous demonstrations, making it comparable to insects in terms of agility.
A survey of UK journalists found that most (56%) use AI weekly, while 27% use it daily. However, concerns remain, with 62% viewing AI as a threat to journalism. The study also revealed a gap between independent and large corporate media outlets in their adoption of AI.
The EBRAINS Summit 2025 will bring together experts to assess how neuroscience can drive medical progress, digital innovation, and responsible data use. Preliminary results from the EPINOV clinical trial, integrating virtual brain technology for epilepsy surgery planning, will be presented.
A researcher warns that AI can distort our perceptions of death and existence by blurring presence and absence. He suggests that embracing selfless selves, or interconnectedness, can help us accept impermanence and the inevitability of death.
The University of Leicester, in partnership with TWI Ltd, has secured £560k to develop a robot-mounted arc-welding system for in-space repair and manufacturing. The project aims to overcome the challenges of welding in space by developing a new, space-qualified robotic welding capability.
Researchers found that complexity can lead to bias in AI models, highlighting the need for better modeling of real-world complexities, input from diverse groups, and ground truths. By addressing these factors, developers can create more accurate and fair AI systems.
A new study by Cambridge University Press & Assessment and Microsoft Research found that traditional learning activities like making notes remain critical for students' reading comprehension and retention. Note-taking, either alone or combined with large language models (LLMs), was more effective in helping students understand and reme...
The article reviews how AI is addressing challenges in metasurface technology, from unit-cell optimization to system-level integration. AI-driven methods like graph neural networks and multi-task learning enhance stability and performance against manufacturing variations.
Researchers at Pusan National University have developed an AI-powered design methodology for gerotor pumps, achieving a 32.3% increase in average flow rate and reducing pressure fluctuation by 53.6%. This innovation has the potential to improve engine durability, lubrication, and cooling, leading to quieter operation and increased reli...
Large language models (LLMs) exhibit network formation behaviors similar to humans, including preferential attachment, triadic closure, and homophily. In real-world simulations, LLMs prioritize homophily most strongly, followed by triadic closure and preferential attachment.