A new ESR-CD model has been proposed to address the student-concept sparsity barrier in cognitive diagnosis. The model uses a unique sparsity-based mask module and achieves significant improvements on various datasets, including AUC gains of 1.5% and 6%.
Researchers identify acetylation as a fine cellular regulatory mechanism that affects pyruvate kinase function and stability. The study reveals isoform-specific regulation of PKM by acetylation, which helps understand how cancer cells proliferate.
Researchers have identified blood-based epigenetic markers that can accurately classify individuals into high-risk prediabetes clusters. These biomarkers reflect different biological signaling pathways and can be used to identify people at high risk of developing type 2 diabetes and its complications early on.
A research team developed BI-TE to protect network topology and bandwidth information from data leakage. The system uses a GNN-based bandwidth utilization prediction model to select optimal forwarding paths, enhancing TE efficiency.
A new Federated Learning defense strategy, LSH-FL, leverages historical gradients to identify malicious clients and mitigate dynamic attack strategies. This approach improves attack resilience while maintaining privacy and model accuracy.
PIVKA-II is a valuable tool for early detection, prognostication, and therapeutic response monitoring in HCC. Its integration into multi-parametric models enhances diagnostic accuracy and enables dynamic risk stratification.
A large longitudinal study found that prolonged physical inactivity increases the body's stress burden by midlife. Engaging in regular physical activity, especially in adulthood, appears to protect against this effect. The study suggests that exercise habits throughout adulthood play a crucial role in coping with long-term stress.
Researchers propose a new presentation plot to report pairwise statistical comparisons of multiple algorithms, avoiding inconsistent results and offering a more intuitive visualization. The study suggests using the Wilcoxon signed-ranks test instead of average ranking strategies.
Hybrid delivery models emerge from practical necessity and improve over time as teams adapt to complex demands. The study reframes hybrid delivery as a dynamic process, highlighting the importance of understanding local conditions and allowing systems to develop iteratively.
The journal has been recognized for its significant international citations, with over 3 million documented in the past year. Its editorial leadership and publishing model ensure rigorous peer review and a multidisciplinary perspective, supporting high-quality research on AI and clinical applications.
Antimicrobial resistance is a fast-growing global threat that could kill more people than cancer by 2050. Scientists discussed data integrity's importance in enabling innovation and adoption during the Frontiers Science House session. A new AGI paradigm grounded in real-time interaction may be needed to overcome current limitations.
A novel SSCIL framework is proposed to leverage unlabeled data for semi-supervised class-incremental learning, improving performance over existing methods. The framework enables gradual acquisition of new class knowledge while maintaining balance between stability and plasticity.
Researchers reconstructed annual canopy height maps for southern China, showing a 61% rise in average canopy height over three decades. Plantation forests grew faster than secondary forests, but ultimately reached greater heights. The study demonstrates the potential of satellite-based monitoring for tracking forest growth and informin...
A new study refines radiocarbon dating of marine remains in the Cantabrian region of Spain, improving precision for reconstructing human past during the Magdalenian period. The research provides new correction values to accurately interpret site chronology and fine-tune the 'clock' used to reconstruct Palaeolithic human populations.
Researchers summarize the role of metabolic reprogramming in driving prostate cancer progression, highlighting key regulators and inter-pathway crosstalk mechanisms. Novel targeted therapeutic strategies targeting glucose, glutamine, and lipid metabolism are proposed to overcome current treatment bottlenecks.
This article analyzes the interplay between multimorbidity and immunosenescence, finding that centenarians can serve as a model of immune resilience. The authors propose using Immune Microenvironment Enhancement Therapy (IMET) to optimize patients' immune systems.
JGU researchers found that specific neurons' synapses lose function with age, leading to a decline in the fly's ability to escape predators. This discovery highlights the trade-off between neural plasticity and vulnerability during aging.
Researchers at Koç University have developed a next-generation, safe, and wireless stimulation technology for retinal degenerative diseases. The system converts near-infrared light into precisely controlled electrical stimulation without causing damage to ocular tissue.
A groundbreaking AI model has achieved unprecedented accuracy in tropical cyclone intensity prediction, marking a significant advancement in weather forecasting technology. The new system, Prithvi-TC, demonstrates superior performance in both accuracy and computational efficiency, particularly in predicting rapid intensification events.
Research from the University of Ottawa finds that a tiny temperature bump lowers nectar quality, resulting in lower body fat reserves for monarch butterflies. The study highlights the subtle threat of climate change on monarch survival.
A two-day oatmeal diet significantly reduced cholesterol levels in participants with metabolic syndrome. The positive effects on cholesterol levels remained stable even after six weeks. The diet also influenced the composition of microorganisms in the gut, contributing to its beneficial effects.
Alo Blac urges data sharing to accelerate cancer precision therapeutics and unlock the disruptive potential of quantum technologies. Researchers, funders, policymakers co-create a forward-looking science roadmap to drive public benefit and private-sector innovation by 2035.
Researchers have developed a photonic Janus dressing that drains, disinfects, and repairs wounds in diabetic mice within 17 days. The membrane uses a hydrophobic-to-hydrophilic gradient to clear exudate and promotes immune reprogramming.
A new LLM-enhanced few-shot entity resolution framework was proposed to improve accuracy and reduce hallucination errors. The FUSER framework achieves higher entity resolution accuracy than existing state-of-the-art approaches, while providing a speedup in uncertainty quantification.
Researchers have integrated AI4DB to improve database optimization, focusing on cardinality/cost estimation, join order selection, end-to-end query optimizers, and text-to-SQL models. These advancements enable more efficient and intelligent database systems for experts and non-experts alike.
Researchers introduced a new approach to grammatical error correction using explanation-based retrieval, improving accuracy and making LLMs more interpretable. By matching sentences with similar error patterns, RE² provides models with relevant examples that directly address mistakes in input.
Epithelial–mesenchymal transition (EMT) contributes to resistance to EGFR-TKIs in non-small cell lung cancer (NSCLC). EMT pathways induced by EGFR-TKIs lead to chemoresistance. Understanding these pathways is crucial for developing targeted treatment approaches.
A randomized clinical trial found that a smart speaker-based intervention significantly improved mental distress, quality of life, and diabetes self-care among older adults with type 2 diabetes. This easily implemented self-management strategy may enhance health outcomes in this population.
FHBL is characterized by low plasma levels of total cholesterol, LDL-C, and APOB, and its pathogenesis involves lipid overload, ER stress, oxidative damage, and autophagic dysfunction. Current challenges include underdiagnosis, sparse epidemiological data, and unclear disease progression mechanisms.
Researchers analyzed 1,696 academic articles and identified five key research clusters, including 'Educational Evaluation and Reform' and 'Talent Cultivation.' The study's findings suggest a shift towards practical, multidimensional evaluation frameworks that consider societal impact and interdisciplinary education.
A new perspective in the Journal of Cognitive Neuroscience explains that unconscious perceptual mechanisms represent an ongoing timeline, which is then consciously remembered. The synthesis of theories suggests that consciousness is a simulation created by the brain, with the purpose of explicit memory to understand the present moment.
A new database, scCASdb, has been developed to standardize single-cell chromatin accessibility sequencing (scCAS) datasets in the h5ad format. This enables diverse single-cell analyses that were previously hindered by the lack of comprehensive collections.
The study reveals that orphan non-coding RNAs (oncRNAs) exist across all major cancer types and can be used as digital molecular barcodes to capture cancer cell identity. The research also identifies oncRNAs that drive cancer progression, with some showing promise as biomarkers for monitoring patients.
A second major neuroscience research project will explore the physiological and psychological impact of shared reading groups and book clubs. The study aims to build a clearer evidence base for how collective experiences shape connection and belonging.
Researchers observed episodic memory in rats, suggesting they can model complex cognitive processes. The study shows rats can recall memories with accuracy and context, even after a long delay, implying evolutionarily ancient forms of memory.
A new study evaluates the accuracy of ten satellite-based and reanalysis precipitation datasets using in-situ rain gauge measurements across Czechia from 2001 to 2021. The GSMaP dataset showed superior performance for rainy days, while ERA5-Land overestimated annual precipitation by 15–35%.
Researchers review low-power memristors for energy-efficient and fast neuromorphic computing. Memristors enable in-memory computing, artificial synapses, and neurons, reducing power consumption. Innovative designs and materials are discussed to overcome challenges.
The Emergency Medical Procedures 3D Dataset (EMP3D) captures intricate movements with unprecedented precision, enabling transformative applications in healthcare AI and robotics. EMP3D offers high-precision reconstruction, AI-ready infrastructure, and open access to accelerate innovation.
This study identifies shared genetic variants linking schizophrenia with left- vs right-hemisphere white-matter tracts, revealing hemispheric asymmetry. Hemispheric specificity was observed in the left posterior thalamic radiation and tapetum, with higher proportions of shared loci.
Researchers developed a nanodelivery system that synergizes ferroptosis and STING activation to treat metastatic bladder cancer. The system showed robust tumor control and improved survival rates, with enhanced immune response and reduced metastasis.
A new study reveals that positively charged ions accumulate in one phase, reducing electrostatic repulsion and allowing negatively charged biomolecules to localize within droplets. This mechanism explains the preferential partitioning of molecules in liquid–liquid phase separation.
The expert consensus provides a structured framework for assessing large language models (LLMs) before deployment in clinical workflows. The framework includes retrospective evaluation, six key capability domains, and essential safeguards for patient data protection and bias mitigation.
Researchers developed bioengineered lymphatic tissue (CeLyT) that restored functional lymph nodes in mice with secondary lymphedema. CeLyTs improved lymphedema symptoms by restoring lymphatic flow, filtration capacity, and immune cell populations.
Insilico Medicine has received investigational new drug (IND) clearance from the US FDA for ISM8969, an orally available NLRP3 inhibitor targeting inflammation and neurodegenerative disorders. The Phase I clinical trial plans to evaluate the safety, tolerability, and pharmacokinetics of ISM8969 in healthy volunteers.
Researchers have found a nonlinear dose-response relationship between exercise and hepatic steatosis, indicating that a combination of aerobic and resistance exercises can produce clinically meaningful improvements. The study suggests that exercising above 130 MET-min/week can lead to significant reductions in hepatic steatosis.
Research reveals significant shifts in tree diversity among tropical forests driven by global change, with some areas experiencing declines in species richness while others gain species.
A study finds that nearly half of newborns with severe spina bifida experience breathing problems during sleep, which can be detected through comprehensive sleep studies. Early treatment may significantly improve cognitive development in these high-risk infants.
Researchers studied diet changes among three whale species, fin, humpback, and minke whales, over 28 years. The study found that whales are adapting their feeding habits in response to limited resource availability and increased competition.
A new study suggests that the Moon's surface record sets a hard limit on the possibility of late meteorite delivery being the dominant source of Earth's oceans. The researchers analyzed oxygen isotopes in lunar regolith samples, finding that even generous assumptions about impactor flux would only supply a small fraction of Earth's water.
Research led by University of Miami researchers shows that effective fisheries management can coexist with predator recovery, contradicting previous assumptions. The study found that fishing effort, not predator recovery, drives fishery yields in the North Sea.
Platelet-derived PITTs (Platelet-derived Integrin- and Tetraspanin-rich Tethers) are released from platelets during severe pathological conditions, driving damaging inflammatory processes. Blocking αIIbβ3 with monoclonal antibodies reduces PITT formation and tissue damage in disease models.
A new machine learning framework for satellite-based CO2 retrievals provides highly accurate concentration estimates while quantifying uncertainty, outperforming traditional physics-based methods in speed and accuracy.
A new study using satellite data confirms that California's shift to electric vehicles is making measurable differences in air quality and public health. For every 200 ZEVs added, nitrogen dioxide levels dropped 1.1%.
Physicists report the first experimental observation of quantum state transfer enabled by hidden symmetries in a network of laser-written optical waveguides. This discovery dramatically expands the design space for quantum circuits, opening the gates towards new classes of networks for secure quantum communication and cryptography.
A new study develops a video-based method to measure cardiopulmonary coupling, synchronizing heartbeat and breathing variations without contact. This approach produces consistent results with traditional devices, offering a promising solution for long-term monitoring and comfort.
A research team at Kumamoto University has developed a breakthrough drug-delivery platform using a small-intestine-permeable cyclic peptide, enabling efficient oral delivery of insulin. This approach achieves a pharmacological bioavailability of approximately 33–41% relative to subcutaneous injection.
Researchers have developed a GNSS-only method that consistently outperforms existing methods in dense urban areas, enabling safer and more reliable autonomous navigation. The approach uses a probabilistic framework to estimate position without relying on carrier-phase integer ambiguity resolution.
Citizen science is essential for improving how restoration success is evaluated and managed over time. By integrating public observations into formal monitoring systems, restored wetlands can be better understood and maintained.
A statistically significant spike in shark bites occurs every October in Hawaiian waters, primarily driven by tiger sharks. The research found a strong link between the October spike and the tiger shark's reproductive cycle, specifically parturition, which occurs during September-October.
The new platform, led by PPPL, aims to speed up simulations needed to advance fusion energy research. STELLAR-AI will integrate CPUs, GPUs, and QPUs to tackle the challenges of private fusion companies, enabling faster design and optimization of stellarator devices.