An Incentive Mechanism for Federated Learning:A Continuous Zero-Determinant Strategy Approach

As a representative emerging machine learning technique, federated learning(FL) has gained considerable popularity for its special feature of “making data available but not visible”. However, potential problems remain, including privacy breaches, imbalances in payment, and inequitable distribution.These shortcomings let devices reluctantly contribute relevant data to, or even refuse to participate in FL. Therefore, in the application of FL, an important but also challenging issue is to motivate as many participants as possible to provide high-quality data to FL. In this paper, we propose an incentive mechanism for FL based on the continuous zero-determinant(CZD) strategies from the perspective of game theory. We first model the interaction between the server and the devices during the FL process as a continuous iterative game. We then apply the CZD strategies for two players and then multiple players to optimize the social welfare of FL, for which we prove that the server can keep social welfare at a high and stable level. Subsequently, we design an incentive mechanism based on the CZD strategies to attract devices to contribute all of their high-accuracy data to FL.Finally, we perform simulations to demonstrate that our proposed CZD-based incentive mechanism can indeed generate high and stable social welfare in FL.

Prediction of lime utilization ratio of dephosphorization in BOF steelmaking based on online sequential extreme learning machine with forgetting mechanism

The machine learning models of multiple linear regression(MLR),support vector regression(SVR),and extreme learning ma-chine(ELM)and the proposed ELM models of online sequential ELM(OS-ELM)and OS-ELM with forgetting mechanism(FOS-ELM)are applied in the prediction of the lime utilization ratio of dephosphorization in the basic oxygen furnace steelmaking process.The ELM model exhibites the best performance compared with the models of MLR and SVR.OS-ELM and FOS-ELM are applied for sequential learning and model updating.The optimal number of samples in validity term of the FOS-ELM model is determined to be 1500,with the smallest population mean absolute relative error(MARE)value of 0.058226 for the population.The variable importance analysis reveals lime weight,initial P content,and hot metal weight as the most important variables for the lime utilization ratio.The lime utilization ratio increases with the decrease in lime weight and the increases in the initial P content and hot metal weight.A prediction system based on FOS-ELM is applied in actual industrial production for one month.The hit ratios of the predicted lime utilization ratio in the error ranges of±1%,±3%,and±5%are 61.16%,90.63%,and 94.11%,respectively.The coefficient of determination,MARE,and root mean square error are 0.8670,0.06823,and 1.4265,respectively.The system exhibits desirable performance for applications in actual industrial pro-duction.

Decoding topological XYZ^(2) codes with reinforcement learning based on attention mechanisms

Quantum error correction, a technique that relies on the principle of redundancy to encode logical information into additional qubits to better protect the system from noise, is necessary to design a viable quantum computer. For this new topological stabilizer code-XYZ^(2) code defined on the cellular lattice, it is implemented on a hexagonal lattice of qubits and it encodes the logical qubits with the help of stabilizer measurements of weight six and weight two. However topological stabilizer codes in cellular lattice quantum systems suffer from the detrimental effects of noise due to interaction with the environment. Several decoding approaches have been proposed to address this problem. Here, we propose the use of a state-attention based reinforcement learning decoder to decode XYZ^(2) codes, which enables the decoder to more accurately focus on the information related to the current decoding position, and the error correction accuracy of our reinforcement learning decoder model under the optimisation conditions can reach 83.27% under the depolarizing noise model, and we have measured thresholds of 0.18856 and 0.19043 for XYZ^(2) codes at code spacing of 3–7 and 7–11, respectively. our study provides directions and ideas for applications of decoding schemes combining reinforcement learning attention mechanisms to other topological quantum error-correcting codes.

Enhancing Image Description Generation through Deep Reinforcement Learning:Fusing Multiple Visual Features and Reward Mechanisms

Image description task is the intersection of computer vision and natural language processing,and it has important prospects,including helping computers understand images and obtaining information for the visually impaired.This study presents an innovative approach employing deep reinforcement learning to enhance the accuracy of natural language descriptions of images.Our method focuses on refining the reward function in deep reinforcement learning,facilitating the generation of precise descriptions by aligning visual and textual features more closely.Our approach comprises three key architectures.Firstly,it utilizes Residual Network 101(ResNet-101)and Faster Region-based Convolutional Neural Network(Faster R-CNN)to extract average and local image features,respectively,followed by the implementation of a dual attention mechanism for intricate feature fusion.Secondly,the Transformer model is engaged to derive contextual semantic features from textual data.Finally,the generation of descriptive text is executed through a two-layer long short-term memory network(LSTM),directed by the value and reward functions.Compared with the image description method that relies on deep learning,the score of Bilingual Evaluation Understudy(BLEU-1)is 0.762,which is 1.6%higher,and the score of BLEU-4 is 0.299.Consensus-based Image Description Evaluation(CIDEr)scored 0.998,Recall-Oriented Understudy for Gisting Evaluation(ROUGE)scored 0.552,the latter improved by 0.36%.These results not only attest to the viability of our approach but also highlight its superiority in the realm of image description.Future research can explore the integration of our method with other artificial intelligence(AI)domains,such as emotional AI,to create more nuanced and context-aware systems.

Dynamic Economic Scheduling with Self-Adaptive Uncertainty in Distribution Network Based on Deep Reinforcement Learning

Traditional optimal scheduling methods are limited to accurate physical models and parameter settings, which aredifficult to adapt to the uncertainty of source and load, and there are problems such as the inability to make dynamicdecisions continuously. This paper proposed a dynamic economic scheduling method for distribution networksbased on deep reinforcement learning. Firstly, the economic scheduling model of the new energy distributionnetwork is established considering the action characteristics of micro-gas turbines, and the dynamic schedulingmodel based on deep reinforcement learning is constructed for the new energy distribution network system with ahigh proportion of new energy, and the Markov decision process of the model is defined. Secondly, Second, for thechanging characteristics of source-load uncertainty, agents are trained interactively with the distributed networkin a data-driven manner. Then, through the proximal policy optimization algorithm, agents adaptively learn thescheduling strategy and realize the dynamic scheduling decision of the new energy distribution network system.Finally, the feasibility and superiority of the proposed method are verified by an improved IEEE 33-node simulationsystem.

Bidirectional rotating direct-current triboelectric nanogenerator with self-adaptive mechanical switching for harvesting reciprocating motion

Amid the growing interest in triboelectric nanogenerators(TENGs)as novel energy-harvesting devices,several studies have focused on direct current(DC)TENGs to generate a stable DC output for operating electronic devices.However,owing to the working mechanisms of conventional DC TENGs,generating a stable DC output from reciprocating motion remains a challenge.Accordingly,we propose a bidirectional rotating DC TENG(BiR-TENG),which can generate DC outputs,regardless of the direction of rotation,from reciprocating motions.The distinct design of the BiR-TENG enables the mechanical rectification of the alternating current output into a rotational-direction-dependent DC output.Furthermore,it allows the conversion of the rotational-direction-dependent DC output into a unidirectional DC output by adapting the configurations depending on the rotational direction.Owing to these tailored design strategies and subsequent optimizations,the BiR-TENG could generate an effective unidirectional DC output.Applications of the BiR-TENG for the reciprocating motions of swinging doors and waves were demonstrated by harnessing this output.This study demonstrates the potential of the BiR-TENG design strategy as an effective and versatile solution for energy harvesting from reciprocating motions,highlighting the suitability of DC outputs as an energy source for electronic devices.

Machine learning applications on lunar meteorite minerals:From classification to mechanical properties prediction

Amid the scarcity of lunar meteorites and the imperative to preserve their scientific value,nondestructive testing methods are essential.This translates into the application of microscale rock mechanics experiments and scanning electron microscopy for surface composition analysis.This study explores the application of Machine Learning algorithms in predicting the mineralogical and mechanical properties of DHOFAR 1084,JAH 838,and NWA 11444 lunar meteorites based solely on their atomic percentage compositions.Leveraging a prior-data fitted network model,we achieved near-perfect classification scores for meteorites,mineral groups,and individual minerals.The regressor models,notably the KNeighbor model,provided an outstanding estimate of the mechanical properties—previously measured by nanoindentation tests—such as hardness,reduced Young’s modulus,and elastic recovery.Further considerations on the nature and physical properties of the minerals forming these meteorites,including porosity,crystal orientation,or shock degree,are essential for refining predictions.Our findings underscore the potential of Machine Learning in enhancing mineral identification and mechanical property estimation in lunar exploration,which pave the way for new advancements and quick assessments in extraterrestrial mineral mining,processing,and research.

Reinforcement learning based adaptive control for uncertain mechanical systems with asymptotic tracking

This paper mainly focuses on the development of a learning-based controller for a class of uncertain mechanical systems modeled by the Euler-Lagrange formulation.The considered system can depict the behavior of a large class of engineering systems,such as vehicular systems,robot manipulators and satellites.All these systems are often characterized by highly nonlinear characteristics,heavy modeling uncertainties and unknown perturbations,therefore,accurate-model-based nonlinear control approaches become unavailable.Motivated by the challenge,a reinforcement learning(RL)adaptive control methodology based on the actor-critic framework is investigated to compensate the uncertain mechanical dynamics.The approximation inaccuracies caused by RL and the exogenous unknown disturbances are circumvented via a continuous robust integral of the sign of the error(RISE)control approach.Different from a classical RISE control law,a tanh(·)function is utilized instead of a sign(·)function to acquire a more smooth control signal.The developed controller requires very little prior knowledge of the dynamic model,is robust to unknown dynamics and exogenous disturbances,and can achieve asymptotic output tracking.Eventually,co-simulations through ADAMS and MATLAB/Simulink on a three degrees-of-freedom(3-DOF)manipulator and experiments on a real-time electromechanical servo system are performed to verify the performance of the proposed approach.

Gas liquid cylindrical cyclone flow regime identification using machine learning combined with experimental mechanism explanation

The flow regimes of GLCC with horizon inlet and a vertical pipe are investigated in experiments,and the velocities and pressure drops data labeled by the corresponding flow regimes are collected.Combined with the flow regimes data of other GLCC positions from other literatures in existence,the gas and liquid superficial velocities and pressure drops are used as the input of the machine learning algorithms respectively which are applied to identify the flow regimes.The choosing of input data types takes the availability of data for practical industry fields into consideration,and the twelve machine learning algorithms are chosen from the classical and popular algorithms in the area of classification,including the typical ensemble models,SVM,KNN,Bayesian Model and MLP.The results of flow regimes identification show that gas and liquid superficial velocities are the ideal type of input data for the flow regimes identification by machine learning.Most of the ensemble models can identify the flow regimes of GLCC by gas and liquid velocities with the accuracy of 0.99 and more.For the pressure drops as the input of each algorithm,it is not the suitable as gas and liquid velocities,and only XGBoost and Bagging Tree can identify the GLCC flow regimes accurately.The success and confusion of each algorithm are analyzed and explained based on the experimental phenomena of flow regimes evolution processes,the flow regimes map,and the principles of algorithms.The applicability and feasibility of each algorithm according to different types of data for GLCC flow regimes identification are proposed.

A dynamic incentive and reputation mechanism for energy-efficient federated learning in 6G

As 5G becomes commercial,researchers have turned attention toward the Sixth-Generation(6G)network with the vision of connecting intelligence in a green energy-efficient manner.Federated learning triggers an upsurge of green intelligent services such as resources orchestration of communication infrastructures while preserving privacy and increasing communication efficiency.However,designing effective incentives in federated learning is challenging due to the dynamic available clients and the correlation between clients'contributions during the learning process.In this paper,we propose a dynamic incentive and reputation mechanism to improve energy efficiency and training performance of federated learning.The proposed incentive based on the Stackelberg game can timely adjust optimal energy consumption with changes in available clients during federated learning.Meanwhile,clients’contributions in reputation management are formulated based on the cooperative game to capture the correlation between tasks,which satisfies availability,fairness,and additivity.The simulation results show that the proposed scheme can significantly motivate high-performance clients to participate in federated learning and improve the accuracy and energy efficiency of the federated learning model.

Exploration on the mechanism of therapeutic and toxic bidirectional effects of Haizao Yuhu decoction based on machine learning and data mining

Objective:To explore the bidirectional mechanism of Haizao Yuhu decoction(HYD)on goiter and drug-induced liver injury(DILI)based on machine learning and data mining.Methods:Firstly,compounds of HYD were selected from the TCMSP,TCMIP,and BATMAN databases,then the TCMSP was used to acquire the targets of compounds.Targets of goiter and DILI were obtained from the GeneCards database.Secondly,common targets of“HYD-goiter”and“HYD-DILI”as well as related compounds were used to construct the networks and perform Random Walk with Restart(RWR)algorithm and network stability test.Finally,core targets in the“HYD-goiter”and“HYD-DILI”networks were used for molecular docking with core compounds and searched for validation on PubChem,and the relevant experimental data of our group were quoted to verify the analysis results.Results:There were 22 intersection targets of HYD and DILI,326 of HYD and goiter.RWR analysis showed that MAPK1,MAPK3,AKT1,etc.may be the core targets of HYD treating goiter,RELA,TNF,IL4,etc.may be the core targets of the bidirectional effect,and eckol may be the core compound in bidirectional effect.Network stability test indicated that the HYD had a high stability on treating goiter and playing a bidirectional effect.The core targets and core compounds docked well,and 37.3%of targets had been confirmed by experiments and 29.8%core targets had been confirmed.Our previous experimental result confirmed that the HYD could treat goiter usefully by reducing the expression levels of PI3K and AKT mRNA,and down-regulating the expression of Cyclin D1 and Bcl-2 mRNA.Conclusion:HYD containing“sargassum-liquorice”combination may have a bidirectional effect on treating goiter and causing DILI.We offered a new way for more explorations on the therapeutic and toxic bidirectional mechanisms based on machine learning and data mining.

Enhancing Deep Learning Soil Moisture Forecasting Models by Integrating Physics-based Models

Accurate soil moisture(SM)prediction is critical for understanding hydrological processes.Physics-based(PB)models exhibit large uncertainties in SM predictions arising from uncertain parameterizations and insufficient representation of land-surface processes.In addition to PB models,deep learning(DL)models have been widely used in SM predictions recently.However,few pure DL models have notably high success rates due to lacking physical information.Thus,we developed hybrid models to effectively integrate the outputs of PB models into DL models to improve SM predictions.To this end,we first developed a hybrid model based on the attention mechanism to take advantage of PB models at each forecast time scale(attention model).We further built an ensemble model that combined the advantages of different hybrid schemes(ensemble model).We utilized SM forecasts from the Global Forecast System to enhance the convolutional long short-term memory(ConvLSTM)model for 1–16 days of SM predictions.The performances of the proposed hybrid models were investigated and compared with two existing hybrid models.The results showed that the attention model could leverage benefits of PB models and achieved the best predictability of drought events among the different hybrid models.Moreover,the ensemble model performed best among all hybrid models at all forecast time scales and different soil conditions.It is highlighted that the ensemble model outperformed the pure DL model over 79.5%of in situ stations for 16-day predictions.These findings suggest that our proposed hybrid models can adequately exploit the benefits of PB model outputs to aid DL models in making SM predictions.

An Underwater Target Detection Algorithm Based on Attention Mechanism and Improved YOLOv7

For underwater robots in the process of performing target detection tasks,the color distortion and the uneven quality of underwater images lead to great difficulties in the feature extraction process of the model,which is prone to issues like error detection,omission detection,and poor accuracy.Therefore,this paper proposed the CER-YOLOv7(CBAM-EIOU-RepVGG-YOLOv7)underwater target detection algorithm.To improve the algorithm’s capability to retain valid features from both spatial and channel perspectives during the feature extraction phase,we have added a Convolutional Block Attention Module(CBAM)to the backbone network.The Reparameterization Visual Geometry Group(RepVGG)module is inserted into the backbone to improve the training and inference capabilities.The Efficient Intersection over Union(EIoU)loss is also used as the localization loss function,which reduces the error detection rate and missed detection rate of the algorithm.The experimental results of the CER-YOLOv7 algorithm on the UPRC(Underwater Robot Prototype Competition)dataset show that the mAP(mean Average Precision)score of the algorithm is 86.1%,which is a 2.2%improvement compared to the YOLOv7.The feasibility and validity of the CER-YOLOv7 are proved through ablation and comparison experiments,and it is more suitable for underwater target detection.

基于Q-Learning的分簇无线传感网信任管理机制

针对无线传感器网络中存在的安全问题,提出了基于Q-Learning的分簇无线传感网信任管理机制(Q-learning based trust management mechanism for clustered wireless sensor networks,QLTMM-CWSN).该机制主要考虑通信信任、数据信任和能量信任3个方面.在网络运行过程中,基于节点的通信行为、数据分布和能量消耗,使用Q-Learning算法更新节点信任值,并选择簇内信任值最高的节点作为可信簇头节点.当簇中主簇头节点的信任值低于阈值时,可信簇头节点代替主簇头节点管理簇内成员节点,维护正常的数据传输.研究结果表明,QLTMM-CWSN机制能有效抵御通信攻击、伪造本地数据攻击、能量攻击和混合攻击.

Machine Learning‑Enhanced Flexible Mechanical Sensing

To realize a hyperconnected smart society with high productivity,advances in flexible sensing technology are highly needed.Nowadays,flexible sensing technology has witnessed improvements in both the hardware performances of sensor devices and the data processing capabilities of the device’s software.Significant research efforts have been devoted to improving materials,sensing mechanism,and configurations of flexible sensing systems in a quest to fulfill the requirements of future technology.Meanwhile,advanced data analysis methods are being developed to extract useful information from increasingly complicated data collected by a single sensor or network of sensors.Machine learning(ML)as an important branch of artificial intelligence can efficiently handle such complex data,which can be multi-dimensional and multi-faceted,thus providing a powerful tool for easy interpretation of sensing data.In this review,the fundamental working mechanisms and common types of flexible mechanical sensors are firstly presented.Then how ML-assisted data interpretation improves the applications of flexible mechanical sensors and other closely-related sensors in various areas is elaborated,which includes health monitoring,human-machine interfaces,object/surface recognition,pressure prediction,and human posture/motion identification.Finally,the advantages,challenges,and future perspectives associated with the fusion of flexible mechanical sensing technology and ML algorithms are discussed.These will give significant insights to enable the advancement of next-generation artificial flexible mechanical sensing.

Transfer learning framework for multi-scale crack type classification with sparse microseismic networks

Rock fracture mechanisms can be inferred from moment tensors(MT)inverted from microseismic events.However,MT can only be inverted for events whose waveforms are acquired across a network of sensors.This is limiting for underground mines where the microseismic stations often lack azimuthal coverage.Thus,there is a need for a method to invert fracture mechanisms using waveforms acquired by a sparse microseismic network.Here,we present a novel,multi-scale framework to classify whether a rock crack contracts or dilates based on a single waveform.The framework consists of a deep learning model that is initially trained on 2400000+manually labelled field-scale seismic and microseismic waveforms acquired across 692 stations.Transfer learning is then applied to fine-tune the model on 300000+MT-labelled labscale acoustic emission waveforms from 39 individual experiments instrumented with different sensor layouts,loading,and rock types in training.The optimal model achieves over 86%F-score on unseen waveforms at both the lab-and field-scale.This model outperforms existing empirical methods in classification of rock fracture mechanisms monitored by a sparse microseismic network.This facilitates rapid assessment of,and early warning against,various rock engineering hazard such as induced earthquakes and rock bursts.

Self-potential inversion based on Attention U-Net deep learning network

Landfill leaks pose a serious threat to environmental health,risking the contamination of both groundwater and soil resources.Accurate investigation of these sites is essential for implementing effective prevention and control measures.The self-potential(SP)stands out for its sensitivity to contamination plumes,offering a solution for monitoring and detecting the movement and seepage of subsurface pollutants.However,traditional SP inversion techniques heavily rely on precise subsurface resistivity information.In this study,we propose the Attention U-Net deep learning network for rapid SP inversion.By incorporating an attention mechanism,this algorithm effectively learns the relationship between array-style SP data and the location and extent of subsurface contaminated sources.We designed a synthetic landfill model with a heterogeneous resistivity structure to assess the performance of Attention U-Net deep learning network.Additionally,we conducted further validation using a laboratory model to assess its practical applicability.The results demonstrate that the algorithm is not solely dependent on resistivity information,enabling effective locating of the source distribution,even in models with intricate subsurface structures.Our work provides a promising tool for SP data processing,enhancing the applicability of this method in the field of near-subsurface environmental monitoring.

Machine learning-guided accelerated discovery of structure-property correlations in lean magnesium alloys for biomedical applications

Magnesium alloys are emerging as promising alternatives to traditional orthopedic implant materials thanks to their biodegradability,biocompatibility,and impressive mechanical characteristics.However,their rapid in-vivo degradation presents challenges,notably in upholding mechanical integrity over time.This study investigates the impact of high-temperature thermal processing on the mechanical and degradation attributes of a lean Mg-Zn-Ca-Mn alloy,ZX10.Utilizing rapid,cost-efficient characterization methods like X-ray diffraction and optical microscopy,we swiftly examine microstructural changes post-thermal treatment.Employing Pearson correlation coefficient analysis,we unveil the relationship between microstructural properties and critical targets(properties):hardness and corrosion resistance.Additionally,leveraging the least absolute shrinkage and selection operator(LASSO),we pinpoint the dominant microstructural factors among closely correlated variables.Our findings underscore the significant role of grain size refinement in strengthening and the predominance of the ternary Ca_(2)Mg_(6)Zn_(3)phase in corrosion behavior.This suggests that achieving an optimal blend of strength and corrosion resistance is attainable through fine grains and reduced concentration of ternary phases.This thorough investigation furnishes valuable insights into the intricate interplay of processing,structure,and properties in magnesium alloys,thereby advancing the development of superior biodegradable implant materials.

Automatic recognition of landslides based on YOLOv7 and attention mechanism

Landslide disasters comprise the majority of geological incidents on slopes,posing severe threats to the safety of human lives and property while exerting a significant impact on the geological environment.The rapid identification of landslides is important for disaster prevention and control;however,currently,landslide identification relies mainly on the manual interpretation of remote sensing images.Manual interpretation and feature recognition methods are time-consuming,labor-intensive,and challenging when confronted with complex scenarios.Consequently,automatic landslide recognition has emerged as a pivotal avenue for future development.In this study,a dataset comprising 2000 landslide images was constructed using open-source remote sensing images and datasets.The YOLOv7 model was enhanced using data augmentation algorithms and attention mechanisms.Three optimization models were formulated to realize automatic landslide recognition.The findings demonstrate the commendable performance of the optimized model in automatic landslide recognition,achieving a peak accuracy of 95.92%.Subsequently,the optimized model was applied to regional landslide identification,co-seismic landslide identification,and landslide recognition at various scales,all of which showed robust recognition capabilities.Nevertheless,the model exhibits limitations in detecting small targets,indicating areas for refining the deep-learning algorithms.The results of this research offer valuable technical support for the swift identification,prevention,and mitigation of landslide disasters.

Intelligent diagnostic model for pterygium by combining attention mechanism and MobileNetV2

AIM:To evaluate the application of an intelligent diagnostic model for pterygium.METHODS:For intelligent diagnosis of pterygium,the attention mechanisms—SENet,ECANet,CBAM,and Self-Attention—were fused with the lightweight MobileNetV2 model structure to construct a tri-classification model.The study used 1220 images of three types of anterior ocular segments of the pterygium provided by the Eye Hospital of Nanjing Medical University.Conventional classification models—VGG16,ResNet50,MobileNetV2,and EfficientNetB7—were trained on the same dataset for comparison.To evaluate model performance in terms of accuracy,Kappa value,test time,sensitivity,specificity,the area under curve(AUC),and visual heat map,470 test images of the anterior segment of the pterygium were used.RESULTS:The accuracy of the MobileNetV2+Self-Attention model with 281 MB in model size was 92.77%,and the Kappa value of the model was 88.92%.The testing time using the model was 9ms/image in the server and 138ms/image in the local computer.The sensitivity,specificity,and AUC for the diagnosis of pterygium using normal anterior segment images were 99.47%,100%,and 100%,respectively;using anterior segment images in the observation period were 88.30%,95.32%,and 96.70%,respectively;and using the anterior segment images in the surgery period were 88.18%,94.44%,and 97.30%,respectively.CONCLUSION:The developed model is lightweight and can be used not only for detection but also for assessing the severity of pterygium.