Donghua University

In the era of deep learning, audio-visual saliency prediction is still in its infancy due to the complexity of video signals and the continuous correlation in the temporal dimension. Most existing approaches treat videos as 3D grids of RGB values and model them using discrete neural networks, leading to issues such as video content-agnostic and sub-optimal feature representation ability. To address these challenges, we propose a novel dynamic-aware audio-visual saliency (DAVS) model based on implicit neural representations (INRs). The core of our proposed DAVS model is to build an effective mapping by exploiting a parametric neural network that maps space-time coordinates to the corresponding saliency values. Specifically, our model incorporates an INR-based video generator that decomposes videos into image, motion, and audio feature vectors, learning video content-adaptive features via a parametric neural network. This generator efficiently encodes videos, naturally models continuous temporal dynamics, and enhances feature representation capability. Furthermore, we introduce a parametric audio-visual feature fusion strategy in the saliency prediction procedure, enabling intrinsic interactions between modalities and adaptively integrating visual and audio cues. Through extensive experiments on benchmark datasets, our proposed DAVS model demonstrates promising performance and intriguing properties in audio-visual saliency prediction.

Smart devices (SDs) used in the Industrial Internet of Things can generate computational tasks for processing the data generated during production. However, due to the limited processing power of SDs, it is necessary to transfer these computational tasks to more powerful devices for processing. To this end, we propose a Mobile Edge Computing (MEC) system based on a Software Defined Network (SDN) for SDs to offload their computational tasks. This MEC system includes multiple MEC servers to handle numerous SDs, which leads to load-balancing challenges among these servers. To tackle this problem, we develop a computational offloading model based on mean-field game theory and introduce a mean-field game-based load-balancing algorithm (MFGLB), which reduces processing latency and facilitates task scheduling through Multi-Agent Deep Reinforcement Learning. Each SD in the MEC system is considered a participant in the mean-field game, simplifying the complex stochastic game into a more manageable dual-agent game. We then prove the existence of Nash Equilibrium for this mean-field game. To evaluate the effectiveness of our MFGLB algorithm, we compare its performance with traditional load-balancing algorithms and a stochastic game-based load-balancing algorithm. Our experimental results demonstrate the superiority of MFGLB in reducing processing latency and addressing load imbalances.

Point cloud registration (PCR) can significantly extend the visual field and enhance the point density on distant objects, thereby improving driving safety. However, it is very challenging for vehicles to perform online registration between long-range point clouds. In this paper, we propose an online long-range PCR scheme in VANETs, called LoRaPCR, where vehicles achieve long-range registration through multi-hop short-range highly-accurate registrations. Given the NP-hardness of the problem, a heuristic algorithm is developed to determine best registration paths while leveraging the reuse of registration results to reduce computation costs. Moreover, we utilize an optimized dynamic programming algorithm to determine the transmission routes while minimizing the communication overhead. To the best of our knowledge, LoRaPCR is the first solution to achieve multi-vehicle point cloud long-range registration. Results of extensive experiments demonstrate that LoRaPCR can achieve high PCR accuracy with low relative translation and rotation errors of 0.55 meters and 1.43 ${}^{\circ }$ , respectively, at a distance of over 100 meters, and reduce the computation overhead by more than 50% compared to the state-of-the-art method.

Detecting cracks is a crucial task to ensure the safety of engineering. However, the complexity of the crack background and the highly uneven pixel ratio between the crack and the background significantly limit the effectiveness of detection algorithms. Therefore, an efficient and innovative algorithm for crack detection is proposed, which utilizes a transformer and a multilevel cross-scale weighted feature fusion module, as well as a progressive transfer learning (TL) strategy. Firstly, in terms of existing algorithms, there are limitations in terms of high leakage and false detection rates, as well as a lack of generalization ability, therefore, a F2N-CrackNet model is proposed, which consists of two efficient feature fusion modules and a Nested Multi-level Attention with Atrous Spatial Pyramid Pooling (NMA-ASPP). Secondly, to tackle sparse and unevenly distributed bridge crack data, enhance model detection capabilities, and expedite convergence, a progressive three-stage hybrid TL strategy incorporating across-domain, inter-domain, and inner-domain transferring is proposed, by migrating knowledge from related domains. Finally, to enrich experimental data supporting and validating the proposed model and the transfer learning approach, a new group of experimental data set is offered, comprising hundreds of high-resolution bridge crack images collecting in Shanghai by using a digital camera. Additionally, an efficient human-computer interactive labeling approach combined with morphological processing is devised. Experiments conducted on both open-source and private datasets reveal substantial performance improvements in crack detection achieved by the proposed model and TL strategy. On private data, the model achieves an 81.28% MIoU, exceeding or approaching the performance of state-of-the-art (SOTA) methods, while simultaneously reducing training time by 5-8 seconds per round compared to SOTA. Meanwhile, the performance on multiple sets of open-source data is similarly elevated.

Electrolyte; Zn‐S Batteries; Dendrite Suppression; Kinetics; Reaction mechanism

The advancement of communication technology has significantly promoted the development of dielectric polymers. However, the quantitative prediction of dielectric constant for rapid material screening is hard due to the low precision of existing theories. In this paper, a new model was developed to calculate the dielectric constants of organic materials through theoretical deduction correlating to the dielectric and polar Hansen solubility parameter (HSP) functions (). This model treated the permanent dipole moments as the primary function determining the dielectric constant and the corresponding polar HSP, which was demonstrated to be in good agreement with experimental data and produced reasonable fitting for organic solvents, thermoplastic polymers, and thermoset polymers, yielding R ² of 0.7488, 0.8104, and 0.7450, respectively, and it demonstrated better fitting with R ² of 0.9043 when applied to organic solvents having low hydrogen bonding component (, < 12.5). This new correlation produces the highest accuracy of prediction when compared to the existing models and provides a better mathematical tool to help design and screen dielectric polymers.

The theory of high entropy‐dissipative structure is confined to high‐entropy alloys and their oxide materials under harsh conditions, but it is very difficult to obtain high entropy‐dissipative structure for smart sensors based on polymers and metal oxides under mild conditions. Moreover, multiple signal coupling effect heavily hinder the sensor applications, and current multimodal integrated devices can solve two signal‐decoupling, but need very complicated process way. In this work, new synthesis concept is the first time to fabricate high entropy‐dissipative conductive layer of smart sensors with triple‐signal response and self‐decoupling ability within poly‐pyrrole/zinc oxide (PPy/ZnO) system. The sensor (SPZ20) amplifies pressure (17.54%/kPa) and gas (0.37%/ppm), reduces humidity (0.41%/% RH) and temperature (0.12%/°C) signals, simultaneously achieving the triple self‐decoupling effect of pressure and gas in the complex temperature‐humidity field because of the enlarged pressure‐contact area, enhanced gas‐responsive sites, altered vapor path and its own heat insulation function. Additionally, it inherits the strong robustness (500 rubbing, washing, and heating or freezing cycles) and endurance (10 000 photo‐purification cycles) of traditional high‐entropy materials for information transmission and smart alarms in emergencies or harsh environments. This work gives a new insight into the multiple‐signal response and smart flexible electronic design from natural fibers.

Electrochemical carbon dioxide reduction to n‐propanol, a high‐energy‐density C3 chemical, presents a promising method for the long‐term storage of renewable electricity. However, the C1‐C2 coupling step, crucial for C3 conversion, suffers from low selectivity and sluggish conversion rate. In this study, a strategy is proposed to regulate the adsorption of C2 active species on Cu by introducing an atomically dispersed Zr, which can effectively enhance the electroreduction of CO2 to n‐propanol. In situ infrared spectroscopy and theoretical studies unveil that the introduce of atomically dispersed Zr modulates the adsorption configuration of *C2 intermediates and strengtnens the binding with *C2 intermediates, thus lowing the energy barrier of the C1–C2 coupling process and accelerating the conversion efficiency. This novel catalyst achieves a n‐propanol Faradaic efficiency of 14.4 ± 0.3% and a high production rate of 70.0 ± 1.0 mA cm⁻², comparable to the best reported values of the CO2‐to‐propanol electroconversion. This study highlights the effectiveness of designing synergistic electrocatalysts to boost the production of high‐value energy products, providing a promising path toward achieving carbon neutrality.

Electrolyte; Zn‐S Batteries; Dendrite Suppression; Kinetics; Reaction mechanism

Complex interactions between the inorganic solid electrolyte (ISE) and the liquid electrolyte (LE) give rise to challenges of achieving durable interface stability in hybrid quasi‐solid electrolytes (HQSE), and the influence on the involved ISE surface ionic conductivity also needs to be investigated. Here, 4‐chlorobenzenesulfonic acid (CBSA) is utilized to establish a self‐assembled monolayer (SAM) on the surface of Li6.4La3Zr1.4Ta0.6O12 (LLZTO), which is then incorporated into PEGDA‐based in‐situ polymerized HQSE. The results show that the introduction of CBSA significantly improves the LLZTO/LE interface stability with the optimized solvation structure, resulting in a favorable ionic conductivity (1.19 mS cm‐1) and an increasing Li+ transference number (0.647). Mechanisms for the promotion of ionic conduction and interfacial stability of SAM‐HQSE are unveiled through the density functional theory (DFT) combined with Raman spectra and 7Li solid‐state nuclear‐magnetic‐resonance. There are no short‐circuits in the Li|SAM‐HQSE|Li cells after 1000 h. The Li|SAM‐HQSE|LFP cells or Graphite|SAM‐HQSE|LFP pouch cells respectively achieve the capacity retention of 91.2% and 87.0% with the 0.5.C‐rate for 500 and 300 cycles. This facile and effective strategy proposed in this work make it accessible for constructing the stable surface micro‐environments of LLZTO where boost and homogenize the Li+ conduction in a hybrid quasi‐solid electrolyte system.

Complex interactions between the inorganic solid electrolyte (ISE) and the liquid electrolyte (LE) give rise to challenges of achieving durable interface stability in hybrid quasi‐solid electrolytes (HQSE), and the influence on the involved ISE surface ionic conductivity also needs to be investigated. Here, 4‐chlorobenzenesulfonic acid (CBSA) is utilized to establish a self‐assembled monolayer (SAM) on the surface of Li6.4La3Zr1.4Ta0.6O12 (LLZTO), which is then incorporated into PEGDA‐based in‐situ polymerized HQSE. The results show that the introduction of CBSA significantly improves the LLZTO/LE interface stability with the optimized solvation structure, resulting in a favorable ionic conductivity (1.19 mS cm‐1) and an increasing Li+ transference number (0.647). Mechanisms for the promotion of ionic conduction and interfacial stability of SAM‐HQSE are unveiled through the density functional theory (DFT) combined with Raman spectra and 7Li solid‐state nuclear‐magnetic‐resonance. There are no short‐circuits in the Li|SAM‐HQSE|Li cells after 1000 h. The Li|SAM‐HQSE|LFP cells or Graphite|SAM‐HQSE|LFP pouch cells respectively achieve the capacity retention of 91.2% and 87.0% with the 0.5.C‐rate for 500 and 300 cycles. This facile and effective strategy proposed in this work make it accessible for constructing the stable surface micro‐environments of LLZTO where boost and homogenize the Li+ conduction in a hybrid quasi‐solid electrolyte system.

Electrocorticography (ECoG) has garnered widespread attention due to its superior signal resolution compared to conventional electroencephalogram (EEG). While ECoG signal acquisition entails invasiveness, the invasive rigid electrode inevitably inflicts damage...

Cancer cells engage in active aerobic glycolysis to meet their bioenergetic synthesis needs, which is known as the Warburg effect. Such a process leads to lactate accumulation in the tumor microenvironment (TME), further promoting cancer progression and inducing immunosuppression. Herein, functionalized dendrimer‐Cu(II) complexes (for short, D‐Cu(II)) as a nanocarrier, which enables a Fenton‐like reaction is developed to generate a large amount of hydroxyl radicals and shows a T1‐weighted magnetic resonance (MR) imaging performance. Importantly, the D‐Cu(II) allows efficient delivery of lactate oxidase (LOx) to cancer cells, directly downregulating the lactate levels. When combining with an immune activator of leukadherin‐1, the developed D‐Cu(II)/LOx complexes alleviate the symptoms of mouse leukemia. With a further coating of macrophage membranes, the generated D‐Cu(II)/LOx@M allows for effective blood‐brain barrier crossing to treat an orthotopic murine glioma model under the guidance of Cu(II)‐facilitated MR imaging. By integrating the LOx‐mediated lactate depletion with Cu(II)‐mediated chemodynamic therapy, the developed dendrimer nanomedicines improve the overall survival and antitumor immune responses of mice, and help to remodel the immunosuppressive TME in both cancer models, greatly sensitizing the treatment efficacy of immune activator. Such dendrimer technology may further be used for theranostics of other cancer types through lactate depletion‐enhanced combinational therapy.

The covalency of the metal─oxygen (M─O) bond is significantly amplified in the transition metal sites with elevated oxidation states, thereby enabling the lattice oxygen‐mediated mechanism (LOM) to transcend the traditional linear scaling limitations of the oxygen evolution reaction (OER). Here, an innovative surface atom release speed‐mediated doping of Mo atoms in NiFe (oxy)hydroxides by controlled dissolution of Mo atoms from the Mo2N surface, which resulted in the formation of NiFeMo(OH)2 with high valence Ni species for OER. Structural characterizations, coupled with in situ Raman and theoretical calculations, elucidate that the incorporation of Mo in NiFeMo(OH)2 modulates the electronic configuration of the metal centers, thereby diminishing the formation energy of Ni 3+/4+ species. This modulation augments the M─O bond covalency, facilitating a shift in the OER pathway from the conventional absorbate evolution mechanism to the more efficient LOM. Consequently, the NiFeMo(OH)2 displays a low overpotential of 236 mV at a current density of 10 mA cm⁻², along with long stability (>500 h) at 50 mA cm⁻². Furthermore, when integrated into an anion exchange membrane water electrolyzer, it achieves a current density of 1.0 A cm⁻² at a cell voltage of merely 2.27 V, underscoring its potential for practical applications.

Multi-sensor information fusion is a technology that collects and processes data from multiple sensors to obtain more comprehensive and accurate information. In the field of robotics, multi-sensor information fusion technology is frequently applied to robots since the use of a single sensor in robots is often limited, this will also improve the accuracy and robustness of robot environment perception. After researching and organizing multi-sensor information fusion technology, robotics technology and other related materials, this paper provides a comprehensive introduction to the application of multi-sensor information fusion technology in the field of robotics, describes the logical structure of multi-sensor information fusion technology in robotics, explains the algorithms of multi-sensor information fusion technology that are mainly applied to robots. Finally, this paper summarizes and exemplifies the main applications of multi-sensor information fusion technology in the field of robotics.

The Chinese government has established definitive goals to reach a "carbon peak" by 2030 and achieve "carbon neutrality" by 2060. Investigating the attainment of these emission reduction objectives while simultaneously fostering regional economic growth and enhancing living standards holds critical importance. This study examines the link between higher education and carbon intensity across China’s thirty provincial-level administrative regions, employing fixed effects models on provincial panel data spanning 2001–2020. The findings, validated through robustness tests and a mediation effect model, elucidate the mechanisms by which higher education influences carbon intensity. Notably, the results reveal that enhancing higher education markedly lowers carbon intensity; specifically, a 1% increase in the logarithmic transformation of per capita investment in higher education in a province decreases its carbon intensity by 0.219%. Additionally, higher education’s output similarly contributes to reductions in carbon intensity. The influence of higher education on reducing carbon intensity is particularly pronounced in the central and western regions of China. Moreover, higher education facilitates the reduction of carbon intensity through mechanisms such as promoting environmental consciousness, advancing industrial structure, and encouraging technological innovation.