Articles tagged with Artificial Intelligence
A study by Cornell University finds that AI-based writing assistants can generate generic language that makes non-Western users write in a more American style, leading to cultural stereotyping and language homogenization. Indian participants showed a smaller productivity boost compared to their American counterparts.
Researchers have developed an AI-driven model that can accurately diagnose various subtypes of pediatric sarcoma using only routine pathology images. The model correctly distinguished between nine sarcoma subtypes, demonstrating its potential to provide quick and accurate cancer diagnoses in under-resourced settings.
A new AI-driven biomarker model analyzes imaging and clinical data to predict cancer cachexia with high accuracy, improving detection rates by up to 85% compared to standard methods. The model's skeletal muscle quantification function was trained using annotated CT scans and validated on a separate set of images.
A recent study used AI to discover that phosphoglycerate dehydrogenase (PHGDH) is a causal gene for spontaneous Alzheimer's disease. The researchers found that altering PHGDH expression levels can significantly impact the progression of the disease, suggesting a novel therapeutic target.
A team of researchers at POSTECH has identified a hidden mechanism in Electrochemical Random-Access Memory (ECRAM) technology, enabling faster and more efficient AI computations. This breakthrough could lead to significant improvements in data processing and reduced energy consumption.
Researchers at Johns Hopkins University found that AI systems struggle to understand social dynamics and context necessary for human interaction. Human participants were able to accurately rate features important for understanding social interactions, while AI models failed to match human brain and behavior responses across the board.
FAU CA-AI will acquire NVIDIA infrastructure for generative physical AI and develop state-of-the-art hardware and software for computational T&E of connected AI autonomy. This funding solidifies FAU's role as a national leader in next-gen networked AI autonomous systems research.
Researchers used AI-driven methods to analyze thousands of digital images of melanoma tumor tissue, identifying key immune cell structures that boost immunotherapy effects. The presence of these structures was linked to significantly better overall survival for patients with advanced melanoma.
A new AI algorithm, Viz HCM, can quickly and specifically identify high-risk heart patients with hypertrophic cardiomyopathy (HCM) and provide individualized risk assessments. The algorithm's findings can help doctors prioritize the highest-risk patients for earlier appointments and treatment.
A new robotic system uses cues in a scene to determine a human's objective and quickly identify relevant objects, enabling intuitive assistance in household, workplace, and warehouse settings. The approach achieved 90% accuracy in predicting human objectives and 96% accuracy in identifying relevant objects.
Researchers at Carnegie Mellon University developed a cybersecurity lab, CyberSim Lab, that integrates experiential learning, role play, and collaborative learning to support learning outcomes. The lab is designed to bridge the gap between classroom education and real-world workplace skills.
A new approach, called ESGAN, reduces the need for human-annotated data by one-to-two orders of magnitude in machine-learning models for crop breeding. The tool can differentiate between aerial images of flowering and nonflowering grasses with minimal external guidance, increasing the pace of agricultural field research.
Stephen S. Morse has been appointed as the new Editor-in-Chief of Disaster Medicine and Public Health Preparedness Journal, with a focus on building knowledge on preventive measures. Dr. Morse will oversee the journal's content, ensuring it translates science into practice and integrates medical and public health perspectives.
A new AI algorithm can automatically assess stent healing in OCT images, differentiating between healthy and unhealthy tissue with high accuracy. This technology has the potential to improve clinical decision-making and reduce healthcare costs by providing precise measurements and valuable insights for patient management.
A study by the University of Zurich found that respondents prioritize present AI risks over hypothetical future threats, highlighting the importance of addressing actual problems. The research suggests a need for concurrent understanding and appreciation of both immediate and potential challenges in public discourse.
Researchers developed a new model called React-OT that can predict the transition state of chemical reactions in under a second with high accuracy. The model uses linear interpolation to generate better initial guesses, reducing the number of steps and computation time needed.
MIT researchers have created a unifying framework that combines existing ideas to improve AI models or create new ones. The 'periodic table of machine learning' categorizes classical algorithms based on the approximate relationships they learn, allowing for fusion of strategies and discovery of new algorithms.
BEACON leverages AI, LLMs, and expert networks to rapidly collect, analyze, and disseminate information on emerging infectious diseases. The platform provides near real-time reports of sentinel cases, clusters, and outbreaks.
The University of Texas at El Paso has launched the AI Institute for Community-Engaged Research (AI-ICER), an interdisciplinary think tank leveraging artificial intelligence technologies to address pressing regional challenges. The institute aims to foster collaboration between researchers, public and private sector groups, and industr...
Researchers develop a new approach to automatically guide large language models (LLMs) towards generating accurate, properly structured outputs for various programming languages and formats. This probabilistic approach boosts computational efficiency, enabling small LLMs to outperform larger models in generating accurate code.
Researchers developed an AI platform that tracks cancer biomarkers to personalize treatment dose adjustments for patients with advanced solid tumors. The study resulted in optimal doses being administered to 97.2% of patients, with average reductions of 20%.
A new AI algorithm has been calibrated to quickly identify patients with hypertrophic cardiomyopathy (HCM) and provide individualized risk assessments. The tool can help prioritize high-risk patients for earlier appointments and treatment, leading to better patient outcomes.
Researchers from UChicago and Google will tackle pressing challenges in AI-generated content detection, privacy protection, and security applications of large language models. The partnership aims to promote responsible AI use and bolster security in an increasingly digital world.
Dr. Rameau's cross-cutting research combines clinical expertise with AI to improve detection and treatment of swallowing disorders in older adults. She has made significant contributions to the field, including developing innovative tools for early diagnosis and intervention.
A newly identified subtype of Castleman disease, oligocentric Castleman disease (OligoCD), has been found to be a distinct clinical entity different from existing classifications. This discovery will help diagnose and properly treat patients who have been caught between classification systems, offering more accurate treatment options.
A recent meta-analysis found that generative AI's diagnostic accuracy is lower than that of specialist doctors, with an average accuracy of 52.1%. The study suggests that while generative AI has the potential to support non-specialist doctors in diagnostics, further research is needed to improve its capabilities.
A survey of patients who underwent breast cancer screening mammograms found most support AI's role as a second reader, but concerns about data privacy and bias persisted. Demographic factors such as education and racial background played a significant role in shaping patient perceptions of AI.
Researchers at Osaka Metropolitan University developed an autonomous driving algorithm for robots to navigate raised cultivation beds, utilizing lidar point cloud data. The system enables precise movement and accuracy in both virtual and actual environments, promising to expand tasks beyond harvesting to monitoring and pruning.
A noise-adapted AI model using lead I ECGs estimated heart failure risk with high accuracy, suggesting a potential strategy for early detection. The study's results highlight the promise of wearable and portable ECG devices in identifying at-risk patients.
MIT researchers have developed a new data-driven method that eliminates redundant computations in complex logistical problems. The approach uses machine learning to predict which operations should be recomputed and reduces the solve time for problems like scheduling trains, hospital staff, and factory tasks.
Binghamton University is launching an Institute for AI and Society with $5 million in New York state funding, enabling researchers to tackle issues like online antisemitism and protecting power systems from malicious attacks. The institute will tap into the power of Empire AI, a consortium of public and private universities in New York.
A new anomaly-based IDS system using machine learning and deep learning models achieves remarkable performance, detecting and classifying complex attacks in real-time. The system's top performer, CatBoost, sets a new benchmark for IoT security with 99.85% accuracy.
Researchers are exploring the use of immunotherapy drugs and a tumor-busting 'oncolytic' virus to tackle high-grade neuroendocrine tumors. A new clinical trial, led by Dr. Aman Chauhan, aims to unlock the mysteries of renal cell carcinoma through detailed laboratory and clinical studies.
Researchers have created a breakthrough photonic chip that can train nonlinear neural networks using light, accelerating AI training while reducing energy use. The chip uses a special semiconductor material to reshape how light behaves, enabling reconfigurable systems with wide mathematical function expression.
Researchers at Max Planck Institute use AI to design novel interferometric gravitational wave detectors, discovering dozens of top-performing designs that surpass known human solutions. These findings have the potential to improve detectable signal range by over an order of magnitude.
A recent study by Concordia researchers uses reinforcement learning to maximize online polarization on Twitter, a platform prone to echo chambers and malicious manipulation. The approach confirms the effectiveness in intensifying disagreements across social networks.
Researchers from Osaka Metropolitan University developed an explainable AI model for ship navigation that explains the basis for its decisions and intentions using numerical values. This technology aims to increase trust among maritime workers and contribute to the realization of unmanned ships.
Researchers found that realistic AI avatars are rated more positively than cartoon-style ones for perceived competence, integrity, and benevolence. However, individual factors such as prior AI knowledge and trust in science moderate perceptions of trustworthiness.
A novel AI tool, RibbonFold, predicts the structures of amyloids, revealing previously overlooked nuances in their formation and evolution. This breakthrough may reshape how researchers approach neurodegenerative disease treatment and offers a scalable method for analyzing harmful protein aggregates.
Dr. Latifur Khan, a renowned computer science professor, has been elected as an AAAS Fellow for his pioneering work in machine learning and big-data analytics. He developed innovative solutions to adapt machine learning models to cybersecurity risks and created an AI-driven tool to analyze political conflict and violence.
Researchers from East China Normal University developed an AI-driven system to analyze classroom videos, revealing teacher-centered instruction prevails in primary and secondary schools. The study also found that older students engage less in critical discussions and more in structured questions.
Scientists have developed an all-optical activation function based on sound waves for photonic computing, enabling the creation of energy-efficient artificial intelligence systems. This breakthrough could potentially facilitate the scaling up of physical computing systems and pave the way for more efficient optical neural networks.
Researchers have developed a pioneering method that combines atomic force microscopy with artificial intelligence to detect changes in cancer cells at a small scale. This enables more accurate and reliable diagnoses, potentially leading to earlier detection and better treatment outcomes.
A new study by Doshisha University demonstrates the feasibility and reliability of fully automated speaking tests for English language learners, enabling more frequent evaluation and larger-scale studies. The approach uses AI-powered speech recognition and computational metrics to assess language proficiency.
Researchers developed a novel 2D phase-transition memristor leveraging intrinsic ion migration to overcome existing device limitations. The device achieves record-low power consumption, ultrafast switching speed, and exceptional endurance, making it suitable for high-speed computing applications.
A new AI-powered training platform is being developed to simulate pipeline incidents and provide real-world scenarios for pipeline operators. The platform will use artificial intelligence to create a realistic environment where teams can practice handling hazardous conditions and responding to emergencies.
Giulio Toscani argues that meditation and critical thinking skills are vital for humans to interact with artificial intelligence responsibly. By reflecting on technology use and its impact, individuals can make informed decisions that enhance the positive impact of AI while mitigating its risks.
Researchers developed a new framework, PAC Privacy, to maintain AI model accuracy and ensure sensitive data remains safe from attackers. The new variant of PAC Privacy estimates anisotropic noise, reducing computational cost and boosting accuracy.
The Deep Computational Text Analyser (DECOTA) is an open-access AI tool that transforms open-ended survey responses into clear themes in minutes, not months. Developed by the University of Bath, DECOTA delivers insights around 380 times faster and over 1,900 times cheaper than human analysis.
A new hardware platform for AI accelerators capable of handling significant workloads with reduced energy requirement has been developed. The platform leverages III-V compound semiconductors to create photonic integrated circuits, which operate at the speed of light with minimal energy loss.
SeamFit's innovative use of flexible conductive threads and machine-learning algorithms accurately detects movements and counts reps during various exercises. This wearable technology promotes practicality in exercise tracking, potentially enhancing human-AI interaction by monitoring daily activities.
Scientists have created a detailed map of U2OS cells, revealing previously unknown protein functions and assemblies. The study will help researchers understand how mutated proteins contribute to childhood cancers and provide a blueprint for mapping other cell types.
Researchers introduce an innovative approach using dynamic topic modeling to analyze annual and quarterly reports, uncovering hidden risk factors. The novel method creates tradable indices with a cost-effective alternative to traditional index construction.
Thoracic surgeons can rely on DeepSeek's accurate diagnoses, precise surgical planning, and early warning systems to enhance patient care. However, challenges persist, such as limited dataset availability and high economic burdens, which must be addressed for seamless integration of AI in thoracic surgery.
Researchers created a promising approach that augments LLMs with graph-based models to generate molecules with desired properties. The method generates molecules with better matching user specifications and valid synthesis plans, outperforming existing LLM-based approaches.
A fully automated ICSI system has successfully conceived the world's first baby, promising to transform fertility treatment with greater standardization and consistency. The system, developed by Conceivable Life Sciences, uses AI to position sperm cells and execute microinjection with unprecedented accuracy.
A recent study from UTSA researchers reveals that large language models (LLMs) can pose a serious threat to programmers who use them to help write code. The study found that up to 97% of software developers incorporate generative AI into their workflow, and 30% of code written today is AI-generated.
A new editorial in Oncotarget discusses how artificial intelligence can improve liver imaging by recognizing when it might be wrong. The approach, called 'uncertainty quantification,' helps clinicians better detect liver cancer and other diseases by pointing out areas in medical scans that need a second look.
A study found that people process movements differently based on avatar appearance, with a neural system dedicated to perceiving others' movements playing a key role. The findings may help scientists improve human-AI interactions.
A new system for early detection of Autism Spectrum Disorder (ASD) has been developed using virtual reality and artificial intelligence. The system, which uses biomarkers related to behavior, motor activity, and gaze direction, achieves an accuracy of over 85%, surpassing traditional methods.