This review summarizes the application of machine learning in material defect research, including defect-data construction, model selection, and validation. It highlights key findings, such as the importance of matching models to defect physics and the need for independent first-principles or experimental validation.
Photoacoustic nondestructive evaluation combines optical absorption with acoustic features to characterize material composition and structures. Representative applications include defect inspection, morphology monitoring, and quantitative material property evaluation in various industries.
A new laboratory study uses large-amplitude oscillatory shear and Fourier-transform rheology to track the shift from solid-like to fluid-like behavior in loess mudflows as deformation increases. The study identifies four strain markers that distinguish the linear regime, nonlinear yielding, and the solid-liquid crossover, with water co...
A deterministic IFC-based workflow evaluates whole-building material choices for embodied carbon and circularity, identifying Pareto-optimal solutions. The workflow uses open BIM data to connect building information with material assessments, providing a traceable basis for exploring early-stage material trade-offs.
The Cambrian explosion's triggers are linked to supercontinent breakup, mantle plume activity, and extreme geomagnetic stress, driving genetic and environmental changes. The multi-sphere driving mechanism coupled these factors to create a habitable environment.
KRAS-driven cancer is a complex disease with multiple layers of post-translational control, including membrane localization, stability, activity, and signaling output. Post-translational modifications of KRAS offer new therapeutic opportunities by targeting PTM enzymes and modification-dependent regulatory networks.
The study establishes a bidirectional trust framework for AI and blockchain interactions, addressing key issues like identity, permission, and intent-based execution. The framework identifies unfinished standards and security questions, highlighting the need for further research to ensure secure and autonomous systems.
Researchers developed a system-level simulation framework to evaluate ultra-dense indoor 5G and 6G wireless networks, integrating 3GPP channel modeling, beamforming, and link adaptation. The framework achieved 42% higher average cell throughput and 160% higher 5th-percentile user throughput under imperfect channel information.
A key-decoupled circuit-model blind quantum computation protocol enhances privacy protection in delegated quantum computing while reducing operational burden. The protocol uses independent encryption and decryption keys to conceal the target operation within its Pauli equivalence class.
A recent study reveals peptide deformylase (PDF) as a novel biomarker for 'cold' Colorectal cancer, correlated with poor prognosis and immune evasion. The study's findings provide a metabolic-immune roadmap for precision immunotherapy, offering a potential target for overcoming immune resistance.
Researchers developed a mechanism-data fusion framework for rapid prediction and optimization of microstructure evolution in laser directed energy deposition of Ti-6Al-4V. The framework uses dual-level long short-term memory networks and multiscale physical modeling, achieving high accuracy and reducing computational cost.
A new machine-learning tool detects illicit stablecoin transactions with high accuracy, helping compliance teams to act on real threats while sparing innocent users. The tool separates distinct types of illicit behavior, such as cybercrime and sanctioned wallets, allowing for more targeted action.
Researchers demonstrate rotary ultrasonic machining can drill high-quality holes in single-crystal silicon by modifying material removal to a hybrid ductile-brittle mode. This process reduces cutting forces by up to 29% while improving surface quality and minimizing tool wear. The findings benefit semiconductor, photovoltaic, and micro...
Researchers propose that wave function phase is the source of randomness in quantum mechanics, potentially leading to new AI advancements. This idea contradicts current understanding and warrants experimental verification.
Researchers develop SIHNO to predict stress fields in porous metamaterials, combining geometric symmetry with Hamiltonian-inspired energy structure. The model outperforms existing approaches, achieving low absolute error and relative L2 error, and maintaining millisecond-level inference speed.
Amino acid sensors play a crucial role in regulating tumor biology by linking nutrient availability to tumor adaptation and immune evasion. Targeting these sensing pathways may expose metabolic vulnerabilities and provide new opportunities for combination cancer therapy.
Researchers developed a filament extrusion system to process recycled PET from waste plastic bottles, producing filament suitable for 3D printing. The recycled PET filament demonstrated competitive mechanical performance compared to PLA, highlighting its potential as a feedstock for additive manufacturing.
Blockchain scalability research proposes a five-layer framework to examine how blockchain can support AI agents, highlighting key challenges in data management, networking, consensus, execution, and application-level mechanisms. The study identifies future directions for building more reliable and scalable AI agent ecosystems.
βIII-tubulin is frequently overexpressed in various cancers and associated with aggressive tumor characteristics and poor therapeutic responses. Targeting βIII-tubulin-associated regulatory networks may provide new strategies to overcome treatment resistance.
Purine metabolism aberrantly reprogrammed in cancer to drive hyperproliferation and metastasis; key enzymes act as active drivers of tumorigenesis. Selective targeting of purine metabolism networks presents promising next-generation anticancer strategy.
A Gipuzkoa study proposes a place-based approach combining digital inclusion, public participation, and institutional coordination to govern AI responsibly. The research highlights the need for territorial digital inclusion to anticipate AI risks and protect vulnerable groups.
Researchers develop AI framework to optimize CMOS LNA designs, reducing power consumption by 62% and improving linearity. The framework increases successful circuit simulations by nearly doubling the rate under foundry design constraints.
A review highlights the potential of reinforcement learning to optimize steel structure design, emphasizing the need for robust code compliance, manageable computational cost, and validated cross-structure generalization. Practical use requires independent engineering verification and testing across clearly defined structural families.
A new study presents high-resolution nitrogen isotope records from the Tarim Basin, documenting a persistent depth-dependent redox gradient across the basin. The findings support the hypothesis that stepwise shallow-ocean oxygenation modulated the multi-phase trajectory of the Cambrian Explosion.
Recent advances in organic photosensitizers for tumor photodynamic therapy enhance reactive oxygen species generation and tumor-killing efficacy. Molecular-structure engineering, nanodelivery systems, and multimodal combination-treatment strategies improve therapeutic outcomes.
Researchers developed an adaptive asymptotic tracking control system for vehicle active suspensions, integrating an event-triggered mechanism and actuator saturation compensation. The technology improves ride smoothness and control accuracy while cutting computational load and communication resource consumption.
Artificial intelligence is enhancing cancer drug discovery by identifying optimal targets, screening compounds, and designing novel therapies. This technology combines various data types to uncover vulnerabilities in cancer cells, ultimately improving the development of effective treatments.
This review highlights metabolic crosstalk between tumor cells and immune cells, shaping the immunosuppressive tumor microenvironment and limiting immunotherapy efficacy. Emerging metabolic intervention strategies aim to overcome therapeutic bottlenecks in skin cancer treatment.
Researchers developed a multi-level decision-making model to select preventive maintenance measures on asphalt pavements. The study prioritized ultra-thin cover and composite seal coat technologies due to their balance between technical performance and economic benefits. The framework provides consistent recommendations for decision-ma...
The study reviews SLM optimization strategies and evaluates a Greek labor-law assistant that combines fine-tuning, Retrieval-Augmented Generation and quantization for practical local deployment. The evaluation used standard language-generation and retrieval-oriented metrics, indicating that larger models achieved stronger scores while ...
Scintillation detectors rely on efficient light collection, and researchers provide a system-level roadmap to improve this efficiency. Key strategies include optimizing crystal geometry, surface treatment, reflectors, optical coupling, photodetector matching, and AI-assisted optimization.
Researchers developed SST-MedChain, a patient-centric framework for secure electronic medical record sharing on permissioned blockchains. The system uses non-interactive delegation, one-time access tokens, and policy-bounded re-delegation to reduce costly on-chain authorization while preserving traceability and revocation.
The study reveals that early-age concrete's reduction in strength originates from the irreversible loss of cohesive strength, while frictional effects become dominant as damage develops. This new experimental method separates cohesive and frictional contributions, providing a clearer way to understand early-age concrete's behavior.
Researchers develop non-Hermitian reshaping engineering to reshape eigenmode profiles without altering the energy spectrum or topology. This approach unlocks a larger design space for reconfigurable devices.
The respiratory microbiota drives lung cancer through four integrated pathways: oncogenic signaling, epigenetic/metabolic reprogramming, chronic immune dysregulation. Microbial signatures can predict diagnosis, staging, and therapy response.
Researchers developed an ensemble deep learning approach to detect cracks and cold flows on aluminum gas meter lids with over 97% accuracy. The system combines three AI models, significantly outperforming traditional manual checks.
Oral squamous cell carcinoma (OSCC) is associated with significant morbidity and mortality due to tumor recurrence, metastasis, and therapeutic resistance. TP53 mutations are strongly linked to advanced tumor stage, lymph node metastasis, therapy resistance, and poor prognosis.
A new framework proposes a third way to balance AI innovation and safety: accelerating responsible innovation through technical, organizational, and ethical advancements. The Telus GenAI customer support agent demonstrates how risks can drive further innovation, reducing the need for restrictive safety constraints.
Researchers developed a data-driven method combining GA-BP neural network and chaotic particle swarm optimization to predict and optimize screen-printing parameters for thick-film resistors. The approach achieved an R² of 0.991, recommending optimal settings in 2.38 seconds and reducing resistance deviation within 5%.
Recent advances in tumor microenvironment-responsive nanomedicines offer a promising strategy to address limitations of traditional cancer therapies. These smart carriers enable precise drug release and enhance therapeutic efficacy through structural transformations triggered by specific stimuli within the tumor microenvironment.
The technology clarifies its core advantages over traditional manufacturing methods and provides a roadmap for future development.
Researchers propose a novel process paradigm of photopolymerization-extrusion coupled molding targeting the 3D printing of polymer-derived ceramics. The technology offers an innovative route for additive manufacturing of complex ceramic components with exceptional properties.
A new transfer learning framework connects two gait analysis tasks, predicting continuous gait cycle percentage and classifying discrete gait phases. The method achieved high F1-scores and efficiency, outperforming traditional approaches.
Researchers developed NSYOLO, an AI framework that automatically segments and analyzes nanoparticles with high precision. The framework achieved a mean Average Precision (mAP@0.5) of 0.957, outperforming baseline models and traditional tools in complex imaging environments.
The resonant trident process is theoretically studied for ultrarelativistic electrons and positrons in a strong circularly polarized wave. The study reveals two characteristic quantum energies, Compton effect energy and Breit-Wheeler energy, which determine the probability of a resonant process.
This review highlights the integration of machine learning with adsorption science and engineering, achieving high precision and interpretability in adsorption processes. The reviewed studies demonstrate that machine learning enables accurate prediction of adsorption performance, accelerates material discovery and process optimization,...
Proteomics-driven precision oncology combines multi-omics integration to reveal tumor heterogeneity and functional regulatory networks. Emerging single-cell and spatial proteomics technologies with AI analysis facilitate clinically relevant biomarker discovery.
A new review highlights the role of intestinal stem cells (ISCs) and their supportive microenvironment in driving colorectal cancer initiation, therapy resistance, and relapse. Targeting the ISC–cancer stem cell ecosystem offers a promising strategy to develop next-generation precision care for patients.
This review highlights the key findings of laser additive manufacturing of metallic lattice structures, including TPMS architectures, node reinforcement, and machine learning-driven optimization. The study demonstrates enhanced mechanical performance, energy absorption, and multifunctional capabilities such as thermal management and vi...
Researchers review advanced AI algorithms and hardware acceleration techniques to predict material properties, optimize structures, and discover new materials. This review provides crucial guidance for accelerating data-driven materials research and fostering next-generation functional materials development.