A study published in Frontiers of Computer Science finds conditions that make a regular balanced random (k,2s)-CNF formula (1,0)-unsatisfiable with high probability. These conditions also apply to random instances of the regular balanced (k-1,2(k-1)s)-SAT problem, providing new challenges for k-SAT problem algorithms.
A new research model proposes a multi-user task migration strategy that balances migration costs and latency. The model achieves fast and stable convergence, outperforming existing methods in terms of system delay and user experience.
The study estimates the temporal and spatial distribution characteristics of livestock and poultry manure production in Shaanxi Province, highlighting areas with high environmental pollution risks. Accurate estimation provides data support for policy formulation and promotes sustainable development of the livestock industry.
Researchers proposed a unified objective for dynamics model and policy learning in model-based reinforcement learning, achieving higher sample efficiency and better performance. The Model Gradient algorithm outperforms existing methods in multiple continuous control tasks, especially in sparse reward tasks.
China's plantation area is under pressure due to reduced afforestation space and quality issues, but has a large carbon sequestration potential. Experts suggest promoting land greening, improving forest management, optimizing plantation structure, and strengthening supervision to address these challenges.
A novel variant in the MYO1D gene is associated with laterality defects, congenital heart diseases, and sperm defects. The study's findings underscore the significance of genetic factors in diagnosing and managing these conditions.
Researchers have developed a new algorithm for fully dynamic all-pairs shortest paths in the Massively Parallel Computation model, achieving lower round complexity and memory usage compared to existing static algorithms. The proposed algorithm combines graph and algebraic methods to reduce computational overhead.
A novel Group Scene Graph Generation method is proposed to understand team sports videos from an athlete's perspective. The approach constructs a hierarchical relation network to establish intra-team and inter-team relationship features, significantly improving performance over baselines.
The research team proposed a general framework named HeterMM, which applies in-DRAM index to heterogeneous memory-based key-value stores on NVM. HeterMM holds hot data in DRAM and optimizes access to read-only data in NVM.
Tregs modulate inflammation to facilitate repair and promote bone formation, enabling improved healing outcomes. The study highlights the potential of targeting immune responses to enhance tissue repair and recovery in fractures.
Researchers developed improved double-layer frame structures with cushion layers to enhance MEA continuity and reduce membrane deformation. This enhances PEMFC durability, crucial for achieving the 5000-hour goal, bringing fuel cell vehicle commercialization closer to reality.
The Gria protocol addresses issues in existing deterministic concurrency control protocols by introducing an auto-scaling batch size, reducing conflicts in low-concurrency scenarios, and eliminating write-after-write conflicts. Experimental results show a significant performance improvement over Aria, outperforming it by 13x.
A research team introduced a new generalized splitting-ring number theoretic transform (GSR-NTT) to accelerate polynomial multiplication. They demonstrated that K-NTT, H-NTT, and G3-NTT can be regarded as special cases of GSR-NTT under different parameterizations, achieving significant speed-ups in lattice-based schemes.
Early detection of Alzheimer's disease is critical due to its progression from normal cognition to dementia. Non-cognitive signs like behavioral symptoms, sleep disorders, and sensory impairments may indicate cognitive decline. The review highlights the potential of blood-based biomarkers and gut microbiome in early diagnosis.
The study investigates the role of OTUB1 in vascular smooth muscle cells during atherosclerosis. Knocking down OTUB1 ameliorates plaque progression and stabilizes atherosclerotic plaques by increasing PDGFRβ stability, inhibiting phenotype switch of VSMCs.
Researchers found that textual features do not surpass nominal features in bug assignment accuracy, even with advanced NLP techniques. Nominal features, such as developer preferences, achieved competitive results without using text.
A new research by Wei Song and team proposes a novel method for explicit data expansion to address challenges in implicit discourse relation classification. The proposed method uses argument pair type classification (APTC) and label-smoothing strategy to filter out noisy argument pairs and reduce the impact of noisy sense labels.
Terc-53 overexpression affects normal aging in mice, leading to cognitive decline and shortened lifespan. Hmmr levels decrease with age, but restoring them improves cognitive abilities and reduces neuroinflammation markers.
Researchers developed a novel approach to monitor perovskite ageing in real-time using terahertz time-domain spectroscopy. This technique allows for the detection of material degradation at specific frequencies, providing an indicator of the ageing degree.
The Adaptive-k method adapts the number of samples selected for updating from mini-batches to effectively separate noisy samples. It outperforms other algorithms on various image and text datasets, demonstrating its potential to revolutionize deep learning model training in label noisy datasets.
Researchers analyzed public perceptions of historic urban landscapes in Shaoxing, China, revealing that different characteristics affect sentiments differently on weekdays and weekends. Urban designers should develop targeted plans to balance preservation with development, ensuring a sustainable environment and people's well-being.
Researchers fabricated dielectric metasurfaces with nanodimples and nanobumps on a flexible polymer substrate, showing controlled transmission and reflection haze across the visible spectrum. This enables increased light absorption in solar cells and LED extraction.
CLIPP monitors optical power by detecting conductance variation caused by surface state absorption, enabling non-invasive on-chip monitoring of large-scale photonic integrated circuits. The technology has been applied to identification and feedback control of optical signals, offering improved stability and performance.
A newly developed community park in Ventura, California, was found to increase physical activity and improve mental well-being among residents in low-income neighborhoods. The study suggests that green spaces play a crucial role in enhancing human health in marginalized areas.
A study on refining waterbird habitat conservation in compact urban areas using Nature-Based Solutions (NbS). The project developed a model with six strategies to restore coastal wetland habitats, resulting in notable increases in target species and ecosystem services. It offers valuable insights for ecological restoration across China.
A nationwide cohort study of 254,670 Chinese adults aged 40-80 years reveals significant interactions between age, gender, and sleep duration on lipid levels. The study highlights the importance of considering age in studies of gender differences in metabolic diseases and advocates for personalized prevention and management strategies.
A recent study found that high nitrogen fertilizer application significantly enhances soil carbon sequestration, primarily through vertical transport of organic matter. This redistribution leads to a greater distribution of new organic carbon in deeper soil layers, resulting in increased soil carbon storage.
Researchers evaluated current carbon sinks in Wensu County, identified spatial patterns, and developed ecological restoration strategies. The study found that southern foothills have the best carbon sequestration capacity, with proposed strategies to boost carbon sinks by integrating with other ecological goals
Researchers developed all-optical routers that guide light based on its wavelength and polarization, achieving efficient optical signal control. These compact devices can handle six types of input light, enhancing information processing capability.
This article discusses co-creative strategies for embracing pyric forces in landscape architecture. It highlights the potential of beneficial fire and promotes collaboration with fire stewards to gain insights. Landscape architects are encouraged to become active stewards themselves.
This article proposes a pathway for creating smart neighborhoods by integrating intelligent technology with scenario-based operations. It explores how this approach can improve the design, construction, and operation of future urban neighborhoods, providing a reference for China's smart neighborhood construction.
ChatGPT has shown impressive capabilities in tasks related to knowledge-mining and text-generation, but exhibited limitations in quantitative analysis and reasoning. The researchers emphasize the importance of human expert augmentation and efficient prompt engineering to minimize hallucination and improve response accuracy.
This study evaluates the quality of GAN-generated results in landscape architecture masterplan generation and their effectiveness in design workflows. Image analysis and user surveys show high levels of visual realism and preference for color and texture in GAN-generated layouts and masterplans.
The Quantitative Biology journal has published a commentary on large cellular models (LCMs), featuring influential authors and AI models such as scBERT and Geneformer. These models have revolutionized single-cell data processing and analysis, offering unprecedented insights into biological processes.
Researchers propose a time-division multiplexing (TDM) planning strategy for compact cities' parking lots, balancing vehicle and pedestrian needs. The approach combines technical expertise with community engagement to create efficient and sustainable spaces.
Researchers developed a new rubber-like optical fiber for UV detection using poly(dimethylsiloxane) doped with an organic dye that acts as a molecular switch. The material can be reused multiple times and is expected to integrate smart textiles and wearable devices for continuous UV dose monitoring.
Site-specific digital twins offer a promising approach to urban design and community engagement by integrating virtual and physical realms, promoting inclusive and responsive development. The concept encourages scalability and specificity in applying DT techniques, facilitating data integration and cross-scale twinning.
The paper reviews existing assessment tools for digitalized low-carbon urban planning and design, highlighting their limitations and application shortcomings. The study categorizes and compares 19 assessment tools, focusing on city-scale and district/neighborhood-scale carbon assessment tools.
Researchers developed an approach to deliver intracellular therapeutic proteins using engineered extracellular vesicles (EVs), bypassing the need for scaffold proteins. This method demonstrated significant activation of interferon signaling and potent antitumor responses in both in vitro and in vivo experiments.
Researchers develop a phytic acid-based nanomedicine that targets and inhibits the mTOR pathway, reducing lipid accumulation and inflammation in liver tissue. The treatment shows promise in alleviating the progression of metabolic dysfunction-associated steatohepatitis (MASH) and improving liver function.
The study found that Ni particle size influences CO2 activation pathways, with smaller particles favoring direct dissociation and larger particles favoring hydrogenation dissociation. Larger particles also show superior resistance to carbon formation.
Researchers developed TurCaMP, a bright cyan fluorescent protein that accurately monitors mitochondrial calcium dynamics. Its unique properties enable stable basal fluorescence in the physiological pH range, facilitating accurate and multiplexed calcium dynamics monitoring.
A study evaluates 21 large language models in mining gene relations and pathway knowledge, finding significant disparities in model performance. API-based models excel in predicting gene regulatory relations, while LLMs show effectiveness in gene network analysis and pathway mapping.
Researchers proposed a WS-SE approach for FAC, incorporating frequency-domain and image-domain features with channel attention, achieving more efficient performance than CNN. The study also introduces a learning strategy from a causal view to exploit attribute label relationships, enabling high-confidence effect attribute inference.
The study evaluated cultivated land quality in China, identifying trends in suitability, contiguity, resilience, and ecological stress. It proposed goals for improving cultivated land quality and constructed a differentiated path to achieve synergies among food security and environmental protection
Saline-alkali land in China offers potential for arable expansion, but requires targeted improvement techniques to reduce salinity and enhance fertility. Researchers propose strategies such as freshwater irrigation, mulched drip irrigation, and crop breeding for saline soil tolerance.
CARM1 plays a significant role in triple-negative breast cancer (TNBC) progression by interacting with HIF1A. The study demonstrates that CARM1 promotes proliferation, invasion, EMT, and stemness in TNBC cells, highlighting its potential as a biomarker for cancer progression.
A new generative model, SCREEN, using masked variational autoencoder and optimal transport mapping outperforms baseline methods in predicting single-cell gene expression responses. SCREEN's robustness to data noise and cell type imbalance makes it applicable to various scenarios.
Researchers aim to develop universal genetic circuits that can be transferred to various organisms, overcame host-specific gene expression machinery and metabolism limitations. Expanding regulatory toolbox and integrating computer-aided design tools will enable automatic design of universal circuits for non-model organisms.
A new research proposes partially-hiding functional encryption schemes with fine-grained access control, enabling ciphertexts to be generated according to an access policy and secret keys associated with attributes and functions. This framework achieves semi-adaptive, simulation-based security under standard assumptions.