Because of computational complexity, the deep neural network (DNN) in embedded devices is usually trained on high-performance computers or graphic processing units (GPUs), and only the inference phase is implemented in embedded devices. Data processed by embedded devices, such as smartphones and wearables, are usually personalized, so the DNN model trained on public data sets may have poor accuracy when inferring the personalized data. As a result, retraining DNN with personalized data collected locally in embedded devices is necessary. Nevertheless, retraining needs labeled data sets, while the data collected locally are unlabeled, then how to retrain DNN with unlabeled data is a problem to be solved. This paper proves the necessity of retraining DNN model with personalized data collected in embedded devices after trained with public data sets. It also proposes a label generation method by which a fake label is generated for each unlabeled training case according to users' feedback, thus retraining can be performed with unlabeled data collected in embedded devices. The experimental results show that our fake label generation method has both good training effects and wide applicability. The advanced neural networks can be trained with unlabeled data from embedded devices and the individualized accuracy of the DNN model can be gradually improved along with personal using.

Hong-Xu Cheng;Le-Tian Huang;Jun-Shi Wang;Masoumeh Ebrahimi

School of Electronic Sicence and Engineering, University of Electronic Science and Technology of China, Chengdu 611731;Beijing Zhaoxin Electronic Technology Co., Ltd, Beijing 100094;School of Electrical Engineering and Computer Science, Royal Institute of Technology, Kista, SE-16440

电子科技学刊

[1]Hong-Xu Cheng;Le-Tian Huang;Jun-Shi Wang;Masoumeh Ebrahimi-.Retraining Deep Neural Network with Unlabeled Data Collected in Embedded Devices)[J].电子科技学刊,2022(01):55-69

Retraining,retrain

Deep,Neural,Network,Unlabeled,Data,Collected,Embedded,Devices,Because,computational,complexity,deep,neural,DNN,embedded,devices,usually,trained,high,performance,computers,graphic,processing,units,GPUs,only,inference,phase,implemented,processed,by,such,smartphones,wearables,are,personalized,model,public,data,sets,may,have,poor,accuracy,when,inferring,retraining,collected,locally,necessary,Nevertheless,needs,while,unlabeled,then,problem,solved,This,paper,proves,necessity,after,It,also,proposes,generation,method,which,fake,generated,each,case,according,users,feedback,thus,can,performed,experimental,results,show,that,our,both,good,effects,wide,applicability,advanced,networks,from,individualized,gradually,improved,along,using

0.481293