Research using machine learning to solve scientific problems. Covers AI quality control, sea-surface temperature estimation, label-free cell analysis, and cyber defense.
Researchers developed a physics-based framework to predict temperature-driven VOC emissions from automotive paint sludge. Higher temperatures increase the release rate of VOCs, with moderate changes leading to substantial increases in quantity and speed of diffusion.
Researchers developed an inverse-design framework to optimize magnonic crystal design, identifying unconventional lattice structures with large band gaps. The approach enables the exploration of previously unexplored material systems and device dimensions, paving the way for high-speed spin-wave computing and energy-efficient devices
Researchers at MIT developed a new technique called VLASH that allows robots to predict their future position, enabling smoother motions and quicker reactions. This breakthrough doubles the speed of robots performing tasks like pick-and-place and boosts performance in dynamic activities.
The review explores how integrating Federated Learning (FL), Reinforcement Learning (RL), and Natural Language Processing (NLP) can overcome modern NLP system limitations, such as protecting user privacy and adapting to changing environments. The study presents a unified framework that combines FL, RL, and NLP as three co-equal pillars.
Researchers have developed an AI tool that can detect online propaganda in Kinyarwanda, a Bantu language spoken by 350 million Africans. The dataset, called KinyaProp, provides examples of misinformation in Kinyarwanda for large language models to learn from and recognize.
Researchers at Duke University have developed a method to systematically develop novel probiotic and prebiotic combinations to maintain gut health and treat gastrointestinal diseases. The approach uses machine learning and automation to explore complex interactions between microbes, nutritional sources, and the environment.
A KAIST research team developed a next-generation world model that learns executable theories from observation alone. The Neural Theorizer (NEO) model discovers reusable primitives and composes them into executable programs to explain new situations.
Researchers at Harvard John A. Paulson School of Engineering and Applied Sciences have developed a new AI framework called Orla that streamlines building and running AI workflows. In tests, Orla reduced computing costs and response times without sacrificing quality.
A team of researchers at Harvard and Max Planck Institute have developed three new functional components for photonic microchips using an inverse design algorithm. The compact designs are about 500 times smaller than conventional designs and offer a path toward higher-performance integrated light technologies.
The health tech industry is evolving with AI-powered wearables that enable real-time data interpretation, reducing centralized infrastructure demands. Pharmaceutical AI tools like NoHarm automate reviews, freeing up resources for medication errors. These innovations transform healthcare, improving care and patient outcomes.
Gert Aarts, a renowned physicist, has been awarded 1.58 million euros in funding from the Wübben Foundation Science to establish an Advanced Professorship at Bielefeld University. He will focus on linking theoretical physics with machine learning and expanding research on strongly interacting matter.
Researchers used AI and single-cell technology to study the 3D genome in brain cells from individuals with Alzheimer's disease. They found increased compartment mingling, reduced gene activity, and altered brain cell organization. The study identifies 3D genome organization as a key layer of Alzheimer's biology.
The City University of New York has received an $18.1 million NSF award to create a cloud-programmable national laboratory that uses artificial intelligence and robotics to speed the discovery, design, and production of advanced bio-inspired materials. Researchers nationwide will have remote access to automated tools for developing sus...
Researchers have developed an AI-powered framework that combines multiple AI technologies with automated experiments to accelerate the discovery of advanced energy materials. The '4th+ paradigm' approach enables near-atomic-level accuracy in predicting material properties and rapidly analyzing experimental data.
A novel cross-modal fusion framework integrates low-altitude drone RSI with ground robot LiDAR-inertial measurement unit (IMU) odometry to create accurate digital models of orchards. The system achieved localization accuracy on the order of a few centimeters, demonstrating robustness to seasonal variations and long-term drift.
Researchers designed an end-to-end workflow to identify new blue OLED materials using AI and quantum chemistry. They developed a virtual library of over 19,000 molecules and used machine learning to select promising candidates, which were then experimentally evaluated and found to have high color purity and efficiency.
A Tulane University team is using AI to discover new superconductors, which could improve the nation's electrical grid, medical imaging, and quantum computing. The project combines high-fidelity calculations, physics-aware AI, and experimental measurements to accelerate discovery.
Researchers used machine learning to analyze thousands of automated experiments and accurately predict how new material compositions will respond to heat, identifying the most promising materials. This approach gives scientists a roadmap for developing more durable perovskite solar cells that can withstand real-world operating conditions.
Researchers developed an AI platform, PeptiVerse, to predict key properties of peptides, enabling early assessment of drug potential. The open-source platform allows users to evaluate ordinary and chemically modified peptides, streamlining the discovery process.
Researchers developed an AI framework for detecting and managing microplastics in wastewater treatment systems. The system uses computer vision and machine learning to predict removal efficiency and identify pollution sources. While AI can complement chemical analysis, major challenges remain before these tools can be widely deployed.
Researchers developed a framework to integrate AI into hospitals, emphasizing patient care, staff experience, and economic sustainability. The Total Mission Value framework aims to ensure high-quality patient care remains the top priority amidst AI's transformative potential.
Exposure to even moderate levels of multiple air pollutants during critical stages of pregnancy may increase the risk of preterm birth. The study found that a mixture of ozone and fine particulate matter posed the strongest relationship to early preterm birth.
A technology has been developed that allows artificial intelligence to inversely determine process conditions for quantum-dot light-emitting diode devices. The technology roughly doubled efficiency and extended operational lifetime more than 40-fold when applied to actual devices.
A new approach combines AI with high-resolution mass spectrometry and toxicology databases to predict biological effects of environmental chemicals. This framework helps researchers prioritize candidates for laboratory testing and health risk assessment.
This volume of SLAS Technology highlights novel laboratory technologies, open-source software, and disease-specific tools for advancing life sciences research and development. The journal emphasizes the importance of education, knowledge exchange, and global community building to drive innovation in biomedical research.
UCSF Health Converge accelerates development of AI tools for real-world care delivery by co-developing solutions with select companies. The program focuses on building patient-centered, clinically effective AI solutions that align with UCSF Health's standards.
Researchers created a tool that leverages large language models to build novel languages with unique grammatical structures and vocabularies. The ConlangCrafter tool generates diverse languages, which can be used for creative applications such as video games and movies, or aid researchers in studying poorly documented languages.
A new study uses machine learning to predict chemical toxicity in rare and endangered species, reducing the need for direct biological testing. The model achieved strong performance predicting acute and chronic toxicity, with life stage being a key factor.
Researchers developed DeepHHF, an AI model that identifies patients at high risk of heart failure up to five years in advance. The model analyzes standard ECG recordings and detects subtle abnormalities that are often imperceptible to the human eye.
Researchers unveil high-performance, earth-abundant Fe-N-C catalysts using AI and quantum chemistry. The 'dual modulation' strategy accelerates oxygen reduction reaction in fuel cells, enabling cheaper and more efficient hydrogen fuel cells.
Researchers developed a hybrid technology combining human learning and machine learning in noninvasive BCIs, producing rapid and sustained gains in motor imagery control. The study demonstrates significant scientific and technological advancement in BCIs, establishing a scalable pathway toward robust neural interfaces.
Researchers created new metal alloys using AI-driven materials design, retaining strength under extreme conditions. The alloys, made of nickel, cobalt, and chromium, outperformed industry standards in properties such as puncture resistance and oxidation resistance.
A novel AI model called BINND has been developed to predict which DNA molecules bind to each other. The model achieved an accuracy of 83.5% in predicting DNA pairs that would bind, surpassing the state-of-the-art model by at least 10%. This improvement has significant utility for biomedical diagnostic tools and DNA computing applications.
Researchers developed a high-resolution monitoring approach combining drone-based multispectral imaging with ensemble machine learning models. The method revealed how surrounding industrial, agricultural, residential, and green areas influence water quality in urban rivers.
A team of researchers from the University of Cambridge and UC Santa Barbara developed 'adversarial' mathematical systems to map out where AI prediction breaks down. They identified two main reasons why machine learning fails: algorithmic limitations and hidden patterns in complex systems.
Researchers developed an auditing technique to test generative AI models for malicious capabilities, without prompting them for illegal outputs. The new approach identifies model variations that can generate CSAM with 100% accuracy.
A research team led by Iowa State University professors is using machine learning and synthetic techniques to discover new magnetic materials with unprecedented properties. The project aims to create ultra-powerful magnets that can improve energy productivity, reduce electricity costs, and enhance American industrial competitiveness.
Peter J. Denning suggests that Turing's stance on artificial general intelligence and the imitation game has led to the AI mess, with a fundamental flaw in understanding tacit knowledge and its representation problem. He argues that machines cannot grasp human emotions, intuitions, or cultural context, making it impossible to achieve h...
Deep learning models accelerate drug design, predict chemical interactions, and engineer stable candidates. AI-powered simulations optimize dosimetry, predicting biodistribution and generating patient-specific digital twins for individualized treatment planning.
A Dartmouth study reveals that people's gaze patterns in new environments contain unique personality preferences. The researchers used eye-tracking data to model individual gaze patterns and create machine-learning models that could distinguish between participants based on their conceptual themes.
Researchers uncovered previously undetected slow slip events in Parkfield, California, and found that these silent fault movements systematically follow increased low-frequency earthquake activity. The discovery suggests that slow slip may play an important role in how stress evolves along active faults.
Researchers at the University of Pennsylvania and Chinese University of Hong Kong created TD3B, an AI framework guiding peptide generation toward candidates predicted to have a desired effect. The tool predicts binding likelihood and determines activation or deactivation of associated cellular machinery.
Researchers found that variable-rate seeding (VRS) can help farmers strike a better balance in corn and soybean fields. However, soybeans proved to be more complicated due to their adaptability to weather conditions. The study aims to make farming more accessible and efficient for small land holders using digital tools and data-driven ...
Researchers have developed a federated learning algorithm that solves the long-standing conflict between robustness and efficiency in AI development. The new approach anonymizes data and reduces single-point failure risks while maintaining speed. By remembering past client interactions, servers can protect against malicious input.
A new study explores an FMQA-based optimization framework for RNA design, revealing that encoding matters in achieving optimal results. The approach identifies high-quality RNA sequence candidates with relatively few evaluations, outperforming competing methods.
Researchers from Penn, NYU, and the Linguistic Data Consortium create virtual patients with adjustable psychiatric symptoms to simulate real-world conversations. The STELLAR platform aims to augment clinician training practices with essential conversation scenarios.
Researchers develop annealable ferromagnetic icosahedral quasicrystals with unprecedented structural quality, revealing intrinsic magnetic properties and magnetic criticality. The discovery enables the first systematic investigations of quasiperiodic magnetism and magnetic criticality in QCs.
The collaboration aims to develop AI-powered ground systems that can assist operators with routine satellite operations, mission scheduling, and data analysis. The partnership seeks to automate routine tasks with human oversight, enabling more efficient management of larger satellite fleets.
A new study links tire pollution to Alzheimer's disease through the exposure to 6PPD-quinone, a chemical formed from shaved-off tire particles. The researchers used computational methods to identify key genes that predict Alzheimer's disease and found strong binding of 6PPD-quinone to these genes.
BetaDescribe, an AI system, converts protein sequences into detailed textual descriptions of their functions and characteristics. The technology helps bridge the gap between characterized and existing proteins in nature, enabling researchers to rapidly generate evidence-based hypotheses regarding unknown proteins.
A new learning-based adaptive tuning method integrates chaotic search with particle swarm optimization to improve stability and solution quality in chaotic search algorithms. The approach consistently achieves better results than conventional methods, providing a practical means of enhancing the performance of chaotic search.
Researchers create tiny swimmers to deliver drugs through the human body, finding they reverse direction in non-Newtonian fluids like mucus and blood. This discovery enhances understanding of fluid behavior and could lead to targeted drug delivery.
A large-scale study of an online patient portal shows that AI-generated responses can introduce errors and extraneous details, leading to increased editing time for physicians. Adapting AI to individual physician communication styles can improve accuracy by 33% and reduce editing by 26%.
Researchers developed an on-chip all-optical supernode for ultra-low-latency deep neural network inference, achieving a 100-fold increase in inference speed while using only one-ninth of computing resources. The system supports high-speed data routing and switching with low loss and flat response over a spectral range exceeding 100 nm.
A study published in PLOS One found that AI-generated impersonations of political debaters were rated as more authentic and relevant by the public than their actual responses. This raises concerns about the potential for targeted misinformation campaigns against specific public figures.
Researchers create custom-fit prosthetic hands with soft magnetic sensors that capture subtle changes in muscle shape and pressure. The system performs consistently and reliably, translating intent into control of a dexterous robotic hand with up to 90% accuracy.
MARVEL has transformed materials research by combining simulations, machine learning, and experiments to predict and design novel materials. Its open-source codes and computational infrastructures have strengthened the field, enabling reproducible and collaborative discoveries.
A new AI model, SpliceSelectNet, accurately predicts RNA splicing by capturing long-range DNA signals. The model's hierarchical Transformer architecture preserves high computational efficiency while maintaining single-nucleotide resolution, enabling accurate analysis of genomic regions.
Researchers use unsupervised learning and neural networks to group anomalous Hall curves into distinct families, creating a roadmap for understanding complex magnetic behaviors. The framework predicts desirable properties and guides the search for rare quantum states.
The review discusses key AI concepts, including algorithms, models, architectures, machine learning, deep learning, and multimodal models. It highlights their clinical applications, such as detection of lymph node metastases, Nottingham grading, biomarker quantification, risk stratification, and prognostic prediction.