Automatic speech recognition(ASR)systems have emerged as indispensable tools across a wide spectrum of applications,ranging from transcription services to voice-activated assistants.To enhance the performance of these...Automatic speech recognition(ASR)systems have emerged as indispensable tools across a wide spectrum of applications,ranging from transcription services to voice-activated assistants.To enhance the performance of these systems,it is important to deploy efficient models capable of adapting to diverse deployment conditions.In recent years,on-demand pruning methods have obtained significant attention within the ASR domain due to their adaptability in various deployment scenarios.However,these methods often confront substantial trade-offs,particularly in terms of unstable accuracy when reducing the model size.To address challenges,this study introduces two crucial empirical findings.Firstly,it proposes the incorporation of an online distillation mechanism during on-demand pruning training,which holds the promise of maintaining more consistent accuracy levels.Secondly,it proposes the utilization of the Mogrifier long short-term memory(LSTM)language model(LM),an advanced iteration of the conventional LSTM LM,as an effective alternative for pruning targets within the ASR framework.Through rigorous experimentation on the ASR system,employing the Mogrifier LSTM LM and training it using the suggested joint on-demand pruning and online distillation method,this study provides compelling evidence.The results exhibit that the proposed methods significantly outperform a benchmark model trained solely with on-demand pruning methods.Impressively,the proposed strategic configuration successfully reduces the parameter count by approximately 39%,all the while minimizing trade-offs.展开更多
With the increasing demand for security,building strong barrier coverage in directional sensor networks is important for effectively detecting un-authorized intrusions.In this paper,we propose an efficient scheme to f...With the increasing demand for security,building strong barrier coverage in directional sensor networks is important for effectively detecting un-authorized intrusions.In this paper,we propose an efficient scheme to form the strong barrier coverage by adding the mobile nodes one by one into the barrier.We first present the concept of target circle which determines the appropriate residence region and working direction of any candidate node to be added.Then we select the optimal relay sensor to be added into the current barrier based on its input-output ratio(barrier weight)which reflects the extension of barrier coverage.This strategy looses the demand of minimal required sensor nodes(maximal gain of each sensor)or maximal lifetime of one single barrier,leading to an augmentation of sensors to be used.Numerical simulation results show that,compared with the available schemes,the proposed method significantly reduces the minimal deploy density required to establish k-barrier,and increases the total service lifetime with a high deploy efficiency.展开更多
基金supported by Institute of Information&communications Technology Planning&Evaluation(IITP)grant funded by the Korea government(MSIT)(No.2022-0-00377,Development of Intelligent Analysis and Classification Based Contents Class Categorization Technique to Prevent Imprudent Harmful Media Distribution).
文摘Automatic speech recognition(ASR)systems have emerged as indispensable tools across a wide spectrum of applications,ranging from transcription services to voice-activated assistants.To enhance the performance of these systems,it is important to deploy efficient models capable of adapting to diverse deployment conditions.In recent years,on-demand pruning methods have obtained significant attention within the ASR domain due to their adaptability in various deployment scenarios.However,these methods often confront substantial trade-offs,particularly in terms of unstable accuracy when reducing the model size.To address challenges,this study introduces two crucial empirical findings.Firstly,it proposes the incorporation of an online distillation mechanism during on-demand pruning training,which holds the promise of maintaining more consistent accuracy levels.Secondly,it proposes the utilization of the Mogrifier long short-term memory(LSTM)language model(LM),an advanced iteration of the conventional LSTM LM,as an effective alternative for pruning targets within the ASR framework.Through rigorous experimentation on the ASR system,employing the Mogrifier LSTM LM and training it using the suggested joint on-demand pruning and online distillation method,this study provides compelling evidence.The results exhibit that the proposed methods significantly outperform a benchmark model trained solely with on-demand pruning methods.Impressively,the proposed strategic configuration successfully reduces the parameter count by approximately 39%,all the while minimizing trade-offs.
基金This research was supported in part by the National Natural Science Foundation of China under Grant Nos.11405145,40241461,61374152,and 61876168Zhejiang Provincial Natural Science Foundation of China under Grant Nos.LY20F020024 and LY17F030016.
文摘With the increasing demand for security,building strong barrier coverage in directional sensor networks is important for effectively detecting un-authorized intrusions.In this paper,we propose an efficient scheme to form the strong barrier coverage by adding the mobile nodes one by one into the barrier.We first present the concept of target circle which determines the appropriate residence region and working direction of any candidate node to be added.Then we select the optimal relay sensor to be added into the current barrier based on its input-output ratio(barrier weight)which reflects the extension of barrier coverage.This strategy looses the demand of minimal required sensor nodes(maximal gain of each sensor)or maximal lifetime of one single barrier,leading to an augmentation of sensors to be used.Numerical simulation results show that,compared with the available schemes,the proposed method significantly reduces the minimal deploy density required to establish k-barrier,and increases the total service lifetime with a high deploy efficiency.
