ABSTRACT
Distributed machine learning (ML) was originally introduced to solve a complex ML problem in a parallel way for more efficient usage of computation resources. In recent years, such learning has been extended to satisfy other objectives, namely, performing learning in situ on the training data at multiple locations and keeping the training datasets private while still allowing sharing of the model. However, these objectives have led to considerable research on the vulnerabilities of distributed learning both in terms of privacy concerns of the training data and the robustness of the learned overall model due to bad or maliciously crafted training data. This article provides a comprehensive survey of various privacy, security, and robustness issues in distributed ML.
ABSTRACT
Effective management of emerging medical devices can lead to new insights in healthcare. Thus, human body communication (HBC) is becoming increasingly important. In this paper, we present magnetic resonance (MR) coupling as a promising method for the intra-body network (IBNet). The study reveals that MR coupling can effectively send or receive signals in biological tissue, with a maximum path loss of PL ≤ 33 dB (i.e. at 13.56 MHz), which is lower than other methods (e.g., galvanic, capacitive, or RF) for the same distance (d = 100 cm). The angular orientation of the transmitter and receiver coils at short and long distances also show a minor variation of the path loss (0.19 ≤ ∆PL ≤ 0.62 dB), but more dependency on the distance (0.0547 dB/cm). Additionally, different postures during the MR coupling essentially does not affect path loss ( ∆PL ≤ ± 0.21 dB). In the multi-nodal transmission scenario, the MR coupling demonstrates that two nodes can simultaneously receive signals with -16.77 dBm loss at 60 cm and 100 cm distances, respectively. Such multi-node MR transmission can be utilized for communication, sensing, and powering wearable and implantable devices.
