A credibility-aware swarm-federated deep learning framework in internet of vehicles

Although Federated Deep Learning(FDL)enables distributed machine learning in the Internet of Vehicles(IoV),it requires multiple clients to upload model parameters,thus still existing unavoidable communication overhead and data privacy risks.The recently proposed Swarm Learning(SL)provides a decentralized machine learning approach for unit edge computing and blockchain-based coordination.A Swarm-Federated Deep Learning framework in the IoV system(IoV-SFDL)that integrates SL into the FDL framework is proposed in this paper.The IoV-SFDL organizes vehicles to generate local SL models with adjacent vehicles based on the blockchain empowered SL,then aggregates the global FDL model among different SL groups with a credibility weights prediction algorithm.Extensive experimental results show that compared with the baseline frameworks,the proposed IoV-SFDL framework reduces the overhead of client-to-server communication by 16.72%,while the model performance improves by about 5.02%for the same training iterations.

Continuous-variable quantum key distribution with on-chip light sources

Integrated quantum key distribution(QKD)systems based on photonic chips have high scalability and stability,and are promising for further construction of global quantum communications networks.On-chip quantum light sources are a critical component of a fully integrated QKD system;especially a continuous-variable QKD(CVQKD)system based on coherent detection,which has extremely high requirements for the light sources.Here,for what we believe is the first time,we designed and fabricated two on-chip tunable lasers for CV-QKD,and demonstrated a high-performance system based on these sources.Because of the high output power,fine tunability,and narrow linewidth,the involved on-chip lasers guarantee the accurate shot-noise-limited detection of quantum signals,center wavelength alignment of nonhomologous lasers,and suppression of untrusted excess noise.The system’s secret key rate can reach 0.75 Mb/s at a 50 km fiber distance,and the secure transmission distance can exceed 100 km.Our results mark a breakthrough toward building a fully integrated CV-QKD,and pave the way for a reliable and efficient terrestrial quantum-secure metropolitan area network.

Decoupling effect and driving factors of carbon footprint in megacity Wuhan,Central China

Background China’s 35 largest cities,including Wuhan,are inhabited by approximately 18%of the Chinese popula-tion,and account for 40%energy consumption and greenhouse gas emissions.Wuhan is the only sub-provincial city in Central China and,as the eighth largest economy nationwide,has experienced a notable increase in energy con-sumption.However,major knowledge gaps exist in understanding the nexus of economic development and carbon footprint and their drivers in Wuhan.Methods We studied Wuhan for the evolutionary characteristics of its carbon footprint(CF),the decoupling relation-ship between economic development and CF,and the essential drivers of CF.Based on the CF model,we quantified the dynamic trends of CF,carbon carrying capacity,carbon deficit,and carbon deficit pressure index from 2001 to 2020.We also adopted a decoupling model to clarify the coupled dynamics among total CF,its accounts,and eco-nomic development.We used the partial least squares method to analyze the influencing factors of Wuhan’s CF and determine the main drivers.Results The CF of Wuhan increased from 36.01 million t CO_(2)eq in 2001 to 70.07 million t CO_(2)eq in 2020,a growth rate of 94.61%,which was much faster than that of the carbon carrying capacity.The energy consumption account(84.15%)far exceeded other accounts,and was mostly contributed by raw coal,coke,and crude oil.The carbon deficit pressure index fluctuated in the range of 8.44-6.74%,indicating that Wuhan was in the relief zone and the mild enhancement zone during 2001-2020.Around the same time,Wuhan was in a transition stage between weak and strong CF decoupling and economic growth.The main driving factor of CF growth was the urban per capita residen-tial building area,while energy consumption per unit of GDP was responsible for the CF decline.Conclusions Our research highlights the interaction of urban ecological and economic systems,and that Wuhan’s CF changes were mainly affected by four factors:city size,economic development,social consumption,and technological progress.The findings are of realistic significance in promoting low-carbon urban development and improving the city’s sustainability,and the related policies can offer an excellent benchmark for other cities with similar challenges.

Butler matrix enabled multi-beam optical phased array for two-dimensional beam-steering and ranging

Based on the wavelength transparency of the Butler matrix(BM)beamforming network,we demonstrate a multibeam optical phased array(MOPA)with an emitting aperture composed of grating couplers at a 1.55μm pitch for wavelength-assisted two-dimensional beam-steering.The device is capable of simultaneous multi-beam operation in a field of view(FOV)of 60°×8°in the phased-array scanning axis and the wavelength-tuning scanning axis,respectively.The typical beam divergence is about 4°on both axes.Using multiple linearly chirped lasers,multibeam frequency-modulated continuous wave(FMCW)ranging is realized with an average ranging error of 4 cm.A C-shaped target is imaged for proof-of-concept 2D scanning and ranging.

Confidence Estimation Transformer for Long-Term Renewable Energy Forecasting in Reinforcement Learning-based Power Grid Dispatching

Expansion of renewable energy could help realize the goals of peaking carbon dioxide emissions and carbon neutralization.Some existing grid dispatching methods integrating short-term renewable energy prediction and reinforcement learning(RL)have been proven to alleviate the adverse impact of energy fluctuations risk.However,these methods omit long-term output prediction,which leads to stability and security problems on optimal power flow.This paper proposes a confidence estimation Transformer for long-term renewable energy forecasting in reinforcement learning-based power grid dispatching(Conformer-RLpatching).Conformer-RLpatching predicts long-term active output of each renewable energy generator with an enhanced Transformer to ensure stable operation of the hybrid energy grid and improve the utilization rate of renewable energy,thus boosting dispatching performance.Furthermore,a confidence estimation method is proposed to reduce the prediction error of renewable energy.Meanwhile,a dispatching necessity evaluation mechanism is put forward to decide whether the active output of a generator needs to be adjusted.Experiments carried out on the SG-126 power grid simulator show that Conformer-RLpatching achieves great improvement over the second best algorithm DDPG in security score by 25.8%and achieves a better total reward compared with the golden medal team in the power grid dispatching competition sponsored by State Grid Corporation of China under the same simulation environment.Codes are outsourced in https://github.com/BUPT-ANTlab/Conformer-RLpatching.

Drosophila Larval Light-Avoidance is Negatively Regulated by Temperature Through Two Pairs of Central Brain Neurons

Dear Editor,The innate preference behaviors of animals can be modified by external environmental conditions.In Drosophila for example,the preference for food and temperature are respectively influenced by the hardness of food and environmental light conditions[1,2].Comparatively,environmental modulation of Drosophila light preference has received less investigation.Drosophila avoids light and prefers darkness in the larval stage[3,4].Drosophila larval photoreceptors,Bolwig's organs,and downstream neurons such as the 5th-lateral neurons[4,5]and the posterior ventral lateral-09 neurons[6],are required for the lightavoidance response.