Research using machine learning to solve scientific problems. Covers AI quality control, sea-surface temperature estimation, label-free cell analysis, and cyber defense.
A new study by MIT researchers introduces a method for machine learning with symmetry that is provably efficient, clarifying a foundational question in the field. This approach could lead to more powerful AI models designed to handle symmetry, benefiting applications such as drug discovery and materials science.
Recent AI-based models, such as Asian Prostate Cancer Artificial Intelligence and Galen Prostate, optimize screening and reduce unnecessary biopsies. These models use multimodal clinical parameters and convolutional neural networks to detect prostate cancer and identify aggressiveness of cancer cells.
The new book highlights the transformative role of artificial intelligence (AI) and machine learning (ML) across various domains, including mechatronics, cybersecurity, digital health, and automation. Readers will gain practical insights into AI-based techniques in power systems, social media management, and healthcare diagnostics.
The journal is seeking high-quality submissions that explore the potential of voice and speech analysis in diverse medical applications. Submissions focus on acoustic and voice analysis techniques, voice or vocal phenotyping, and vocal biomarkers in health and medicine.
This book provides a beginner-friendly resource on the impact and evolution of decentralized networks, highlighting their applications in healthcare, supply chains, agriculture, climate monitoring, and education. The authors emphasize sustainability, data security, and ethical tech adoption.
Researchers at Chan Zuckerberg Biohub San Francisco and Stanford University have developed an AI-driven Virtual Lab that can tackle sophisticated scientific problems. The team of AI agents, each equipped with varied expertise, can formulate, refine, and carry out complex research strategies, producing results that can be validated in r...
Researchers developed a new machine learning method to improve climate model projections, making them more accurate at both global and regional levels. The new tool addressed challenges in capturing observed patterns in compound extreme events, leading to improved projections of these events.
Researchers used AI to predict key properties of halide perovskites, enabling rapid design and optimization of more efficient solar materials. The findings provide valuable insights into the rational design of halide perovskites with tailored properties.
The report outlines five key considerations for data transformation in primary care, including automation of data collection and integration with human workflows. Effective implementation requires cross-sectoral collaborations between government, industry, and academia to upgrade human and data infrastructures.
A new study reveals that pedestrians are now walking faster and spending less time in public spaces. Researchers analyzed 40 years of video footage to find a 14% decline in people lingering in these areas.
A team of computer scientists created 2,300 original sudoku puzzles and asked AI tools like OpenAI's ChatGPT to solve them. The results showed that while some AI models could solve easy sudokus, most struggled to provide accurate explanations, raising questions about the trustworthiness of AI-generated information.
Researchers used machine learning to detect subtle behavioral changes in mouse models of Parkinson's disease, revealing that high-speed movements are the first affected behaviors. Levodopa treatment improves movement speed but not other attributes of these movements.
Researchers at the University of Rochester are developing biologically inspired predictive coding networks for digital image recognition using analog circuits, which could lead to more efficient drones. The team aims to approach the performance of existing digital approaches and translate it to complex perception tasks needed by self-d...
Researchers have developed new coarse-grained (CG) models to simulate ionic liquids (ILs), addressing their high viscosity in molecular dynamics simulations. These models offer deeper insights into IL structure-property relationships and enable applications in biological and electrochemical systems.
A new framework developed by Northwestern University and UCLA scientists integrates various water-related processes with a machine-learning model to predict landslide threats. The framework identifies three main pathways leading to landslides: intense rainfall, rain on already saturated soils, and melting snow or ice.
Researchers developed a new AI model named BATMAN to improve T cell receptor therapy accuracy. The AI uses a vast database of over 22,000 TCR-peptide interactions to predict peptide binding and identify potential cancer treatments.
Answer ALS is launching a groundbreaking collaboration with Tulane University and the Pennington Biomedical Research Center to harness AI for ALS treatment discovery. The Louisiana AI Drug Development Infrastructure for ALS (LADDIA) will prioritize therapeutic targets using AI-driven insights from the Answer ALS' Neuromine Data Portal.
The team aims to deliver AI power directly to devices, improving resilience and speed in constrained environments. By processing data step-by-step across a network of devices, they can create a safe and adaptable system that can withstand attacks and extreme conditions.
Researchers use AI to solve differential equations, such as Schrodinger's equation, for large-scale systems, improving efficiency and accuracy in fields like drug discovery and material design.
Researchers developed an AI-based screening tool using pangrams to detect Parkinson’s disease with nearly 86 percent accuracy. The web-based test analyzes voice recordings for subtle patterns linked to the neurodegenerative disease, identifying potential warning signs.
Researchers developed an AI-based method to analyze microscope images of corroded metal surfaces, estimating corrosion severity without human input. The technique focuses on two critical indicators: corrosive deposit thickness and porosity, identifying specific pH levels that indicate severe corrosion stages.
Researchers from Florida Atlantic University's Harbor Branch Oceanographic Institute have developed a multi-sensor biologging tag on the whitespotted eagle ray, enabling long attachment times and detailed data collection. The study provides insights into the rays' behavior, habitat use, and social interactions.
A study found that even advanced AI models can default to intuitive answers, overlooking critical details in health care settings. Researchers tested commercially available LLMs using creative lateral thinking puzzles and modified well-known medical ethics cases, revealing blind spots that clinicians can't afford.
Researchers at UMBC developed a new way to predict 2D materials that could transform the electronics industry. Using a mix of data mining, computer modeling, and structural analysis, they predicted 83 possible new materials with desirable properties.
A University of Kansas study found that people rate corporate crises messages written by humans as more credible and trustworthy, regardless of the approach taken. However, the approach itself didn't vary between participants who read human or AI-written content.
A new UNLV-led study uses AI to detect emerging virus variants in wastewater samples, outperforming existing methods. The algorithm can identify unique signatures for different virus variants with as few as two to five samples, significantly earlier than current methods.
MIT engineers developed a versatile demonstration interface that allows users to teach robots new skills in three intuitive ways: remote control, physical manipulation, or demonstration. This innovation expands the type of users and 'teachers' who interact with robots, enabling robots to learn a wider set of skills.
The review proposes a systematic AI-driven framework for OLED material design, offering theoretical guidance and practical pathways for intelligent materials development. Machine learning models are used to predict core optoelectronic properties, enabling rapid and accurate property prediction while providing structural insights that g...
Researchers found that deep neural networks exhibit absorbing phase transitions, a phenomenon observed in physical systems like forest fires. This discovery provides a unified framework describing how the signal propagates between layers of neurons, enabling prediction of trainability and generalizability.
CodeSteer, a smart coach developed by MIT researchers, guides large language models (LLMs) to switch between text and code generation to solve complex problems. The system boosts the accuracy of LLMs on symbolic tasks like scheduling shipments in supply chains and multiplication problems.
Scientists use human-AI collaboration to tackle complex questions in condensed matter physics, leveraging machine learning algorithms to identify patterns in simulation data. This approach successfully models the behavior of frustrated magnets and sheds light on quantum computing and gravity.
Researchers from the Weizmann Institute of Science and Intel Labs have developed new algorithms that allow AI developers to combine the power of different AI models, speeding up large language model performance by 1.5-2.8 times. This enables faster collaboration between models, reducing processing power costs and improving overall effi...
Researchers developed OmicsTweezer, a tool that uses machine learning and single-cell data integration to analyze human tissue. The tool can estimate cell type composition in tumors and surrounding tissues, which could help pinpoint potential therapeutic targets.
Researchers created an 'encyclopedia of sugar utilization pathways' in 263 Bifidobacterium genomes. This resource helps predict which strains thrive in different conditions, enabling personalized probiotics to match children's lifestyles and dietary needs. The study suggests a potential solution to improving outcomes for preterm infants.
A newly developed AI-based tool uses protein expression to create a stemness index that analyzes tumor similarity to pluripotent stem cells. The model predicts tumor aggressiveness and identifies potential targets for new therapies, offering hope for improved cancer treatment.
Two Bar-Ilan University researchers have won ERC grants to advance cancer treatments using targeted protein degradation and improve regulatory effectiveness through machine learning. The projects aim to pioneer new therapeutics and design more effective interventions based on empirical evidence.
A new study from UNC-Chapel Hill reveals repetitive flooding in North Carolina is far more common and widespread than previously recognized. Over 90,000 buildings flooded in at least one event, with 43% located outside of FEMA's designated high-risk areas.
A new study by Cornell University reveals Amazon's AI shopping assistant Rufus gives vague or incorrect responses to users writing in some English dialects like AAE. The researchers propose a framework for evaluating chatbots that can better serve users from diverse linguistic backgrounds.
Researchers found that friends tend to perceive products similarly and exhibit synchronized brain activity when viewed together, even predicting each other's purchasing intentions. As friendships deepen, this similarity strengthens.
The UK Biobank has completed a groundbreaking imaging project, scanning the brains, hearts, abdomens, blood vessels, bones and joints of 100,000 volunteers to understand how diseases develop. The project is already improving patient care in the NHS and beyond, with over 1,300 peer-reviewed scientific papers published based on the data.
Researchers developed a self-driving lab that collects at least 10 times more data than previous techniques, dramatically expediting materials discovery research while slashing costs and environmental impact. The system uses dynamic flow experiments to continuously characterize samples, capturing data every half second.
A recent study utilizes AI to optimize catalyst design and synthesis, predicting structure-property relationships and minimizing resource-intensive calculations. The research also explores the integration of automated synthesis, characterization, and optimization in closed-loop systems.
Researchers developed an AI system that enables a four-legged robot to adapt its gait to different terrain, just like animals. The robot learned to switch gaits on the fly and navigate uneven surfaces without any alterations to the system itself, overcoming previous limitations around adaptability.
The journal is seeking original research, viewpoints, and tutorials exploring the use of AI-driven documentation in clinical settings. Submissions should address practical applications, challenges, and future directions of AI-facilitated clinical documentation.
Researchers developed an AI model that accurately predicts genetic mutations from routine pathology slides, reducing the need for rapid genetic testing by over 40% in lung adenocarcinoma patients. This innovation streamlines clinical decision-making and accelerates access to targeted therapies.
A novel AI-based approach identifies a distinct physicochemical signature near the cleavage site of gamma-secretase substrates, revealing dynamic properties essential for molecular recognition. The study highlights the potential of this methodology to improve understanding of gamma-secretase's role in diseases and aid drug development.
Researchers developed a smarter QKD system using machine learning to synchronize quantum signals, achieving reliable self-correction in unstable timing. This breakthrough enables compact and cost-effective quantum communication systems for secure key sharing over long distances.
Researchers developed a way to boost LLMs' performance on challenging tasks using test-time training. This method involves updating some model parameters using new data, leading to significant improvements in accuracy, especially for tasks requiring logic and reasoning.
A recent study using machine learning and computational modeling reveals that astrocytes play a more active role in brain function than previously thought. Astrocytes subtly modulate communication between neurons during synchronous brain activity, influencing network coordination and stability.
Researchers developed Group Encoding (GE) to forecast electricity demand using On/Off device status, improving prediction accuracy and efficiency in smart energy operation. The method simplifies complex datasets while retaining key information for optimal energy management.
A new study reveals that AI systems transition from relying on word positions to meaning-based understanding as they receive enough data for training. The transition occurs abruptly, similar to a phase transition in physical systems, and is driven by the amount of data available.
Researchers have launched an open-source platform to benchmark, improve and run competitions for single-cell genomics computational methods. The platform standardizes evaluations, fosters reproducibility and accelerates progress towards open challenges in this fast-moving field.
A new technique called WeGeFT improves large language model performance in tasks such as commonsense reasoning and code generation. By fine-tuning key parameters, researchers reduce the need for significant computational power, advancing the field of artificial intelligence.
Researchers developed an AI-powered microscope system to measure soil fungi presence and quantity, providing insights into soil health and fertility. The low-cost optical microscopy with machine learning technology can be used by farmers and land managers worldwide.
A fully autonomous robotic system developed by MIT researchers can measure important material properties like photoconductivity, increasing the speed and precision of research. The system uses machine learning and robotics to analyze new semiconductors and optimize the development of more powerful solar panels.
The College of Engineering at Texas A&M is developing a suite of university-wide resources to integrate generative AI into course material, research, and outreach. The initiative aims to make generative AI a core part of the academic toolkit accessible to faculty across disciplines.
A team of researchers at Tohoku University's AIMR used machine learning potential to characterize Sn catalyst activity, identifying the most effective catalysts for CO2 reduction. The study provides novel insights into the behavior of Sn-based catalysts and could lead to more efficient fuel production.
New research suggests that blurry, color-limited vision in early childhood can contribute to the development of key brain pathways. The study proposes that such limited vision allows the brain to specialize in certain processing units, which later develop into distinct pathways for color and fine spatial detail.
The Hebrew University of Jerusalem is part of a €9.3 million, 36-month Horizon Europe project to develop trustworthy GenAI systems. The consortium aims to integrate neural and symbolic methods for human-centric design in complex systems.
Researchers at Helmholtz Munich developed an AI model called Centaur that can simulate human behavior with remarkable accuracy. It was trained on over ten million decisions from psychological experiments and makes decisions that closely resemble those of real people. The model opens new avenues for understanding human cognition and imp...