Federated Learning enables privacy-preserving training of Transformer-based language models,but remains vulnerable to backdoor attacks that compromise model reliability.This paper presents a comparative analysis of de...Federated Learning enables privacy-preserving training of Transformer-based language models,but remains vulnerable to backdoor attacks that compromise model reliability.This paper presents a comparative analysis of defense strategies against both classical and advanced backdoor attacks,evaluated across autoencoding and autoregressive models.Unlike prior studies,this work provides the first systematic comparison of perturbation-based,screening-based,and hybrid defenses in Transformer-based FL environments.Our results show that screening-based defenses consistently outperform perturbation-based ones,effectively neutralizing most attacks across architectures.However,this robustness comes with significant computational overhead,revealing a clear trade-off between security and efficiency.By explicitly identifying this trade-off,our study advances the understanding of defense strategies in federated learning and highlights the need for lightweight yet effective screening methods for trustworthy deployment in diverse application domains.展开更多
Aiming at the sensitivity of linear active disturbance rejection control(LADRC)to measurement noise,an improved anti-saturation cascaded LADRC is proposed.This approach employs the system output as the control input o...Aiming at the sensitivity of linear active disturbance rejection control(LADRC)to measurement noise,an improved anti-saturation cascaded LADRC is proposed.This approach employs the system output as the control input of the filtering subsystem,which is then fed back to the secondary LADRC to mitigate measurement noise and uncertain disturbances.While preserving the benefits of precise and stable tracking inherent to traditional cascaded LADRC closed-loop systems,this design omits the outer loop tracking differentiator,thereby simplifying the structure and reducing the number of tunable parameters.Additionally,an error compensation strategy is introduced to address input saturation constraints,thereby equipping the controller with anti-saturation capabilities.Under multi-track surface switching conditions,the effectiveness of the new cascaded active disturbance rejection method is verified by simulation of the optimal adhesion control of the railway train.The results show that the improved cascaded LADRC has stronger rapidity and robustness.展开更多
基金supported by a research fund from Chosun University,2024.
文摘Federated Learning enables privacy-preserving training of Transformer-based language models,but remains vulnerable to backdoor attacks that compromise model reliability.This paper presents a comparative analysis of defense strategies against both classical and advanced backdoor attacks,evaluated across autoencoding and autoregressive models.Unlike prior studies,this work provides the first systematic comparison of perturbation-based,screening-based,and hybrid defenses in Transformer-based FL environments.Our results show that screening-based defenses consistently outperform perturbation-based ones,effectively neutralizing most attacks across architectures.However,this robustness comes with significant computational overhead,revealing a clear trade-off between security and efficiency.By explicitly identifying this trade-off,our study advances the understanding of defense strategies in federated learning and highlights the need for lightweight yet effective screening methods for trustworthy deployment in diverse application domains.
基金supported by the State Key Laboratory of Heavy-duty and Express High-power Electric Locomotive open project(Grant No.QZKFKT2023-012).
文摘Aiming at the sensitivity of linear active disturbance rejection control(LADRC)to measurement noise,an improved anti-saturation cascaded LADRC is proposed.This approach employs the system output as the control input of the filtering subsystem,which is then fed back to the secondary LADRC to mitigate measurement noise and uncertain disturbances.While preserving the benefits of precise and stable tracking inherent to traditional cascaded LADRC closed-loop systems,this design omits the outer loop tracking differentiator,thereby simplifying the structure and reducing the number of tunable parameters.Additionally,an error compensation strategy is introduced to address input saturation constraints,thereby equipping the controller with anti-saturation capabilities.Under multi-track surface switching conditions,the effectiveness of the new cascaded active disturbance rejection method is verified by simulation of the optimal adhesion control of the railway train.The results show that the improved cascaded LADRC has stronger rapidity and robustness.
