A survey on deploying mobile deep learning applications:A systemic and technical perspective

With the rapid development of mobile devices and deep learning,mobile smart applications using deep learning technology have sprung up.It satisfies multiple needs of users,network operators and service providers,and rapidly becomes a main research focus.In recent years,deep learning has achieved tremendous success in image processing,natural language processing,language analysis and other research fields.Despite the task performance has been greatly improved,the resources required to run these models have increased significantly.This poses a major challenge for deploying such applications on resource-restricted mobile devices.Mobile intelligence needs faster mobile processors,more storage space,smaller but more accurate models,and even the assistance of other network nodes.To help the readers establish a global concept of the entire research direction concisely,we classify the latest works in this field into two categories,which are local optimization on mobile devices and distributed optimization based on the computational position of machine learning tasks.We also list a few typical scenarios to make readers realize the importance and indispensability of mobile deep learning applications.Finally,we conjecture what the future may hold for deploying deep learning applications on mobile devices research,which may help to stimulate new ideas.

Defect engineered nickel hydroxide nanosheets for advanced pseudocapacitor electrodes

While the past years have witnessed great achievement in pseudocapacitors,the inauguration of electrode materials of high-performance still remains a formidable challenge.Moreover,the capacity and rate capability of the electrode depends largely on its electrical conductivity,which ensures fast charge transfer kinetics in both the grain bulk and grain boundaries.Here,nickel hydroxides with oxygen vacancies are facilely fabricated via a hydrothermal method.The active materials exhibit a high specific capacitance of 3250 F·g^(−1)and a high areal of capacitance of 14.98 F·cm^(−2)at 4.6 mA·cm^(−2).The asymmetric supercapacitor device based on our material delivers a high energy density of~71.6 Wh·kg^(−1)and a power density of~17,300 W·kg^(−1)and could retain~95%of their initial capacitance even after 30,000 cycles.In addition,the defect-rich hydroxides demonstrate higher electrical conductivity as well as dielectric constant,which is responsible for the superior pseudocapacitive performance.Our new scientific strategy in terms of taking the advantages of oxygen vacancies might open up new opportunities for qualified pseudocapacitive materials of overall high performances not only for nickel hydroxides but also for other metal hydroxides/oxides.

Mn-based cathode materials for rechargeable batteries

The rapid expansion of renewable energies asks for great progress of energy-storage technologies for sustainable energy supplies,which raises the compelling demand of high-performance rechargeable batteries.To satisfy the huge demand from the coming energy-storage market,the resource and cost-effectiveness of rechargeable batteries become more and more important.Manganese(Mn)as a key transition element with advantages including high abundance,low cost,and low toxicity derives various kinds(spinels,layered oxides,polyanions,Prussian blue analogs,etc.)of high-performance Mn-based electrode materials,especially cathodes,for rechargeable batteries ranging from Li-ion batteries,Na-ion batteries,aqueous batteries,to multivalent metal-ion batteries.It is anticipated that Mn-based materials with Mn as the major transition-metal element will constitute a flourishing family of Mn-based rechargeable batteries(Mn RBs)for large-scale and differentiated energy-storage applications.On the other hand,several critical issues including Jahn-Teller effect,Mn dissolution,and O release greatly hinder the pace of Mn RBs,which require extensive material optimizations and battery/system improvements.This review aims to provide an investigation about Mn-based materials and batteries for the coming energy-storage demands,with compelling issues and challenges that must be overcome.

Enhancing electrical conductivity of single-atom doped Co_(3)O_(4) nanosheet arrays at grain boundary by phosphor doping strategy for efficient water splitting

High electrical conductivity guarantees a rapid electron transfer and thus plays an important role in electrocatalysis.In particular,for the single atom catalysts(SACs),to facilitate interaction between the single atom and supports,precisely engineering the conductivity represents a promising strategy to design SACs with high electrochemical efficiency.Here we show rhodium(Rh)SAC anchored on Co_(3)O_(4) nanosheets arrays on nickel foam(NF),which is modified by a facile phosphorus(P-doped Rh SACCo_(3)O_(4)/NF),possessing an appropriate electronic structure and high conductivity for electrocatalytic reaction.With the introduction of P atom in the lattice,the electrocatalyst demonstrates outstanding alkaline oxygen evolution reaction(OER)activity with 50 mA·cm^(−2) under overpotential of 268 mV,6 times higher than that of Ir/C/NF.More interestingly,the P-doped Rh SAC-Co_(3)O_(4)/NF can get 50 mA·cm^(−2) at only 1.77 V for overall water splitting.Both electrical conductivity studies and density functional theory(DFT)calculations reveal that the high conductivity at grain boundary improves the charge transfer efficiency of the Rh catalytic center.Furthermore,other noble-metal(Ir,Pd,and Ru)doped Co_(3)O_(4) nanosheets arrays are prepared to exhibit the general efficacy of the phosphorus doping strategy.

Laser processed micro-supercapacitors based on carbon nanotubes/manganese dioxide nanosheets composite with excellent electrochemical performance and aesthetic property

Micro-supercapacitors with excellent electrochemical performance and aesthetic property are realized using the carbon nanotubes/manganese dioxide nanosheets（CNTs/δ-MnO2） composite as electrodes.This CNTs/d-MnO2 nanocomposite is excellently compatible with the slurry dispensing process for electrode fabrication, and thus is conducive for preparing thick electrode films, which exhibits a specific capacitance of 257 F/g with an electrode thickness of 13μm. By involving laser-scribing technique, the electrode film can be patterned with a high resolution and fabricated into a planar micro-supercapacitor,showing the maximum energy density of 6.83 mWh/cm^3 at the power density of 154.3 mW/cm^3, and maintained a value of 2.71 mWh/cm^3 at the maximum power density of 2557.5 mW/cm^3. Considering the versatility of the laser-scribing technical platform, the micro-supercapacitors fabricated in this way exhibit excellent aesthetic property and can cater to various miniaturized wearable electronic applications. This technology opens up opportunities for facile and scalable fabrication of high performance energy devices with shape diversity and a meaning of art.

Decision making framework for autonomous vehicles driving behavior in complex scenarios via hierarchical state machine

With the development of autonomous car,a vehicle is capable to sense its environment more precisely.That allows improved drving behavior decision strategy to be used for more safety and effectiveness in complex scenarios.In this paper,a decision making framework based on hierarchical state machine is proposed with a top-down structure of three-layer finite state machine decision system.The upper layer classifies the driving scenario based on relative position of the vehicle and its surrounding vehicles.The middle layer judges the optimal driving behavior according to the improved energy efficiency function targeted at multiple criteria including driving efficiency,safety and the grid-based lane vacancy rate.The lower layer constructs the state transition matrix combined with the calculation results of the previous layer to predict the optimal pass way in the region.The simulation results show that the proposed driving strategy can integrate multiple criteria to evaluate the energy efficiency value of vehicle behavior in real time,and realize the selection of optimal vehicle driving strategy.With popularity of automatic vehicles in future,the driving strategy can be used as a reference to provide assistance for human drive or even the real-time decision-making of autonomous driving.