Privacy concerns arise from malicious attacks on Deep Neural Network (DNN) applications during sensitive data inference on edge devices. Membership Inference Attack (MIA) is developed by adversaries to determine whether sensitive data is used to train the DNN applications. Prior work uses Trusted Execution Environments (TEEs) to hide DNN model inference from adversaries on edge devices. Unfortunately, existing methods have two major problems. First, due to the restricted memory of TEEs, prior work cannot secure large-size DNNs from gradient-based MIAs. Second, prior work is ineffective on output-based MIAs. To mitigate the problems, we present a depth-wise layer partitioning method to run large sensitive layers inside TEEs. We further propose a model quantization strategy to improve the defense capability of DNNs against output-based MIAs and accelerate the computation. We also automate the process of securing PyTorch-based DNN models inside TEEs. Experiments on Raspberry Pi 3B+ show that our method can reduce the accuracy of gradient-based MIAs on AlexNet, VGG-16, and ResNet-20 evaluated on the CIFAR-100 dataset by 28.8%, 11%, and 35.3%. The accuracy of output-based MIAs on the three models is also reduced by 18.5%, 13.4%, and 29.6%, respectively.

Our work extends Juba’s formulation of learning abductive reasoning from examples, in which both the relative plausibility of various explanations, as well as which explanations are valid, are learned directly from data. We extend the formulation to consider partially observed examples, along with declarative background knowledge about the missing data. We show that it is possible to use implicitly learned rules together with the explicitly given declarative knowledge to support hypotheses in the course of abduction. We observe that when a small explanation exists, it is possible to obtain a much-improved guarantee in the challenging exception-tolerant setting.

Recently, ridesharing applications, such as Uber, have gained popularity by helping riders to save money. However, current real-time ridesharing systems with time-window constraints have not yet considered the impact of real-time road conditions. In this work, we propose transferring algorithms and incentive models to enhance the flexibility of an existing large scale real-time ridesharing system. Transferring algorithms aim to transfer those users who experienced delay to other vehicles by using a pair wise bounding-box pruning technique. The incentive models avoid the use of transferring but instead compensate affected passengers with credits to maintain their satisfaction level in a different way. Extensive experiments were conducted using a real road network. Results show that our proposed algorithms and models are able to boost users' satisfaction level up to 90% under the influence of real-time road conditions.