Researchers discovered a link between ppGpp and sleep regulation in fruit flies, showing that ppGpp is involved in the connection between sleep and starvation. Mutations in Mesh1 gene affected ppGpp degradation, impacting sleep patterns.
The study reveals a widespread presence of Bathyarchaeia species in coastal sediments, which encode novel methyltransferases utilizing lignin-derived aromatics. These enzymes facilitate O-demethylation and have significant implications for the marine ecosystem's function and carbon cycling.
Researchers designed PdFe/Cu nanocatalysts using a stepwise co-reduction method, exhibiting enhanced oxidation properties and durability. The addition of Cu to the precursor transformed spherical PtFe nanoparticles into chains with improved methanol oxidation capabilities.
This study investigates large language model (LLM) construction, optimization, and evaluation, highlighting the importance of open-source models and cost-saving methods. The authors also identify challenges faced by LLMs, including scarcity of datasets and model instability, and propose potential research directions.
The study found that MdMYB73 activates MdPH5 expression and enhances its activity, regulating malate transport into vacuoles. This results in increased malate accumulation and vacuolar acidification.
This study investigates the DNA methylation profiles of Verticillium dahliae and its contribution to fungal pathogenicity. The researchers found that DNA methylation is essential for the penetration and colonization of V. dahliae in plants by inhibiting stress-responsive protein kinase VdRim15.
Researchers developed a method to generate large numbers of homozygous mutant barley plants by treating microspores with ethyl methanesulfonate. This approach provides an efficient way to produce novel genetic resources for pre-breeding and theoretical research.
Nickel-based catalysts have shown high efficiency in CO2 reduction, producing valuable materials through various methods. Recent advancements focus on improving catalyst design and systems, with gas diffusion electrodes enhancing current density.
Researchers evaluate Google Bard's visual comprehension capabilities through 15 scenarios, identifying strengths and limitations in tasks like object detection, recognition, and fine-grained analysis. Bard excels in some areas but struggles with others, highlighting room for improvement in its visual understanding.
A new transgenic maize population shows high glyphosate resistance after co-expression of the GAT and GR79-EPSPS genes. The resulting event, GG2, demonstrates low glyphosate residues and improved agronomic traits.
A recent special issue investigates the molecular mechanisms behind cancer's malignant phenotype, exploring topics such as metabolic reprogramming and tumor heterogeneity. The findings aim to contribute to the development of mechanism-based approaches for cancer treatment.
This review summarizes six common RNA modifications, including m6A, m5C, m1A, m7G, Ψ, and A-to-I editing, which play crucial roles in various human diseases. The authors discuss the mechanisms of these modifications and their targeting by small molecule inhibitors.
The SafePASS MR Training Tool uses Mixed Reality to simulate LSA lifeboat operations, providing crew members with hands-on practice and real-time feedback. The application has been successfully tested and validated in lab and real-world environments, demonstrating improved training outcomes and enhanced awareness of equipment usage.
This study proposes an improved A* algorithm for AGV task path planning, focusing on complex environments and indoor applications. The new algorithm optimizes expansion mode, turning points, and smoothness, reducing unnecessary turns and improving efficiency.
Prof. Chilai Chen's team successfully tested the first deep-sea mass spectrometer of China in the deep sea. This technology enables real-time detection of dissolved gases, shedding light on marine biogeochemical cycles, biodiversity, and global climate change.
Researchers propose a novel architecture, multi-agent decision transformers (MADT), to solve complex decision-making tasks in multi-agent reinforcement learning. By leveraging sequence modelling and attention mechanisms, MADT achieves fast adaptation and superior performance via learning one big sequence model.
This paper proposes an FSSiamese target tracking algorithm based on Digital Twins network, building a GeoAI experimental platform to analyze urban video data. The proposed framework enhances geospatial perception, understanding, and decision-making in urban Digital Twins.
Ni-based catalysts have shown promising performance in the electrooxidation of methanol in alkaline media due to their anti-poisoning ability and high oxidation kinetics. The development of new Ni-based catalysts with varied structures has improved catalytic performance, offering a more sustainable alternative to traditional noble meta...
The Adaptive Spatio-Temporal Attention Neural Network (ASTANN) is proposed for cross-database micro-expression recognition. It extracts optical flow information and combines it with facial images to generate new representations, which are then processed by a deep neural network with spatiotemporal attention mechanisms.
The proposed multi-scale facial video pulse extraction network uses separable spatiotemporal convolution and dimension separable attention to extract pulse signals from facial videos. This approach yields better results than single-scale methods, especially when fusing multiple scales.
Researchers identified ZapE as a key regulator of Pseudomonas aeruginosa biofilm formation, modulating the pqs quorum sensing system. The study suggests that ZapE is a novel target for anti-Pseudomonas compounds and has implications for treating antibiotic-resistant infections.
Researchers identified a Pcc1-like protein in archaeon Sulfolobus islandicus, suggesting it's a homolog of eukaryotic Gon7. The study suggests archaeal KEOPS complex functions independently of t6A modification and supports the archaeal origin of Eukaryotes theory.
Pre-training in medical data tackles challenges such as data scarcity and privacy concerns using AI techniques like transfer learning and self-supervised learning. Recent advances and new frontiers of pre-training-based techniques are introduced, including applications to medical images, bio-signal data, EHR data, and multi-modality data.
A novel photocatalytic reactor uses TiO2 nanotube arrays coupled with nanobubbles to efficiently degrade organic pollutants like Rhodamine B. The method shows high degradation efficiency and can be applied to treat wastewater.
Porphyrinic metal-organic frameworks (PMOFs) have been synthesized using various methods, including solvothermal synthesis. The native biocompatible properties of porphyrins endow PMOFs with great potential in biological applications. Several possible frameworks can be formed with a large scale of metal ions.
The study proposes a spectral feature-based compression method for vibrotactile acceleration signals, which achieves high time-frequency similarity and a 10% compression ratio. The method is suitable for long-term data compression and can further reduce the compression ratio for multi-interaction points.
The IoT-based fire information system utilizes a 3D Geographic Information Science (GIS) platform to visualize fires and facilitate rescue planning. The system incorporates an FMDM (fire monitoring and decision-making) system with features like equipment status, alarm classification, and evacuation path optimization.
A new video anomaly detection algorithm (COVAD) uses content-based attention to focus on objects in frames, improving performance over baseline models. The algorithm also refines the memory module for normal behavioral patterns.
This study introduces a novel approach to image aesthetic assessment using deep learning, focusing on color composition and space formation. It proposes extracting color palette features and image contour maps to evaluate image aesthetics.
A brain-inspired intelligent robotic system is constructed to deal with current limitations, introducing core neural mechanisms and corresponding algorithms. The simulation platform integrates brain-inspired algorithms in vision, decision-making, and movement control, providing efficient tools for researchers.
Researchers developed view interpolation networks to reproduce material appearances of specular objects using images from intermediate viewpoints of four cameras on a sphere. The networks outperformed traditional methods, but future studies are needed to address complex materials and increased degrees of freedom.
Researchers propose URDRN to effectively recover image details and extract accurate rain information. The model outperforms state-of-the-art methods in both subjective and objective evaluation metrics for synthetic and real rain images.
Recent studies using first-principles calculations and micromagnetic simulations have shown that chiral magnetic domain walls can coexist with the quantum anomalous Hall effect in specific materials, such as VSe2 and Fe2XI. This control enables precise manipulation of dissipationless chiral edge states.
Researchers developed a sensitive and selective electrochemical sensor using AuPd@Fe2O3 nanoparticles to detect dopamine. The sensor shows high sensitivity and anti-interference capabilities, making it suitable for detecting dopamine in pharmaceuticals and biological samples.
Researchers developed a method for creating cobalt/carbon nanocomposites using magnetic induction heating, which significantly enhances the oxygen evolution reaction performance. The new synthesis method produces high-performance electrocatalysts with low overpotentials and high current densities.
Researchers have made significant progress in developing high-performance Pt-based nanocatalysts for oxygen reduction reactions, focusing on increasing activity per site and active sites. The study aims to reduce the cost of PEMFCs by overcoming the challenges associated with Pt metal catalysts.
Researchers introduced a framework to support human collaborative intelligence in digital twins using augmented reality. The HCLINT-DT framework showed good adaptability levels in various use cases, including family photo albums and industrial electrical engineering contexts.
An edge-based digital twin service was evaluated for remote control of robotic arms, revealing that inverse kinematics computation and movement trajectory planning are critical functions. The service profile also shows potential for computational resource savings with low-abstraction level commands.
The study proposes an MEC-based framework to improve data delivery, energy consumption, and security in IIoT systems. The proposed solution enhances scalability, reliability, and real-time control, providing a more efficient communication environment.
This study proposes using gamification and virtual reality to create quality learning experiences for digital twin students and employees. The technology aims to enhance learner motivation, acquire knowledge, and foster a personalized and cooperative learning environment.
The article discusses the development of the metaverse through digital twins, emphasizing their focus on simulating real-world processes and entities. Digital twins are virtual replicas that synchronize data between the physical and virtual worlds, while the metaverse focuses on human interaction and connection.
This study proposes DSD-MatchingNet, a novel network leveraging sparse-to-dense hypercolumn matching and deformable convolutional networks for robust local feature matching. The proposed framework generates multilevel dense feature maps and incorporates pixel-level correspondence estimation error to improve accuracy.
The NPIPVis system combines data visualization and machine learning to analyze NBA game data, predicting player performance and team outcomes. By integrating multiple techniques, such as PAOHvis and iStoryline, users can gain a deeper understanding of the data and make informed decisions.
A new gesture drawing algorithm uses deep learning to produce rough, loose sketches with an overdrawn appearance. The algorithm can be controlled using parameters such as stroke length and curvature, enabling various artistic styles.
The study developed a VR application for immersive multi-user firefighter-training scenarios, addressing disaster management tasks with minimal risk. The prototype integrated VR motion capture and synchronization of sensor data to create realistic training environments.
This study evaluates the uncanny valley theory through human interactions with four human-like entities: a voice assistant, child-sized robot, virtual human, and life-sized humanoid robot. The results show that anthropomorphic robots are most liked, while the 'uncanny valley' effect is not observed.
RADepthNet separates depth-related features from irrelevant information and incorporates boundary features for better depth prediction results. The proposed method achieves state-of-the-art performance in monocular depth estimation on two datasets, including a newly built soccer video dataset.
This study uses ultra-pH-sensitive polymeric fluorescent probes to image local pH changes in electrocatalytic processes with subsecond resolution. The technique allows for rapid imaging of proton concentration changes, enabling the study of heterogeneous reactions and mass transport in electrocatalysis.
This study introduces a novel virtual nasal endoscopy system based on computed tomography scans, allowing for realistic training and pathologies simulation. The proposed approach enables efficient development of new training sets, optimizing 3D models while maintaining graphic quality.
The study presents UnityMol as a successful software tool for VR porting, enabling enhanced digital twin models. Interactive simulations also have great potential for research and collaborative problem-solving.