Variational Information Bottleneck Model for Accurate Indoor Position Recognition

Affiliation auteurs!!!! Error affiliation !!!!
TitreVariational Information Bottleneck Model for Accurate Indoor Position Recognition
Type de publicationConference Paper
Year of Publication2021
AuteursQian W, Gechter F
Conference Name2020 25TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR)
PublisherInt Assoc Pattern Recognit; IEEE Comp Soc; Italian Assoc Comp Vis Pattern Recognit & Machine Learning
Conference Location10662 LOS VAQUEROS CIRCLE, PO BOX 3014, LOS ALAMITOS, CA 90720-1264 USA
ISBN Number978-1-7281-8808-9
Résumé

Recognizing user location with WiFi fingerprints is a popular approach for accurate indoor positioning problems. In this work, our goal is to interpret WiFi fingerprints into actual user locations. However, WiFi fingerprint data can be very high dimensional in some cases, we need to find a good representation of the input data for the learning task first. Otherwise, using neural networks will suffer from severe overfitting. In this work, we solve this issue by combining the Information Bottleneck method and Variational Inference. Based on these two approaches, we propose a Variational Information Bottleneck model for accurate indoor positioning. The proposed model consists of an encoder structure and a predictor structure. The encoder is to find a good representation in the input data for the learning task. The predictor is to use the latent representation to predict the final output. To enhance the generalization of our model, we also adopt the Dropout technique for each hidden layer of the decoder. We conduct the validation experiments on a real-world dataset. We also compare the proposed model to other existing methods so as to quantify the performances of our method.

DOI10.1109/ICPR48806.2021.9412651