Guangquan Xu

Guangquan Xu, GuoHua Xin, Litao Jiao et al.

With the growing popularity of Android devices, Android malware is seriously threatening the safety of users. Although such threats can be detected by deep learning as a service (DLaaS), deep neural networks as the weakest part of DLaaS are often deceived by the adversarial samples elaborated by attackers. In this paper, we propose a new semi-black-box attack framework called one-feature-each-iteration (OFEI) to craft Android adversarial samples. This framework modifies as few features as possible and requires less classifier information to fool the classifier. We conduct a controlled experiment to evaluate our OFEI framework by comparing it with the benchmark methods JSMF, GenAttack and pointwise attack. The experimental results show that our OFEI has a higher misclassification rate of 98.25%. Furthermore, OFEI can extend the traditional white-box attack methods in the image field, such as fast gradient sign method (FGSM) and DeepFool, to craft adversarial samples for Android. Finally, to enhance the security of DLaaS, we use two uncertainties of the Bayesian neural network to construct the combined uncertainty, which is used to detect adversarial samples and achieves a high detection rate of 99.28%.

Shangming Cai, Dongsheng Wang, Haixia Wang et al.

To reduce uploading bandwidth and address privacy concerns, deep learning at the network edge has been an emerging topic. Typically, edge devices collaboratively train a shared model using real-time generated data through the Parameter Server framework. Although all the edge devices can share the computing workloads, the distributed training processes over edge networks are still time-consuming due to the parameters and gradients transmission procedures between parameter servers and edge devices. Focusing on accelerating distributed Convolutional Neural Networks (CNNs) training at the network edge, we present DynaComm, a novel scheduler that dynamically decomposes each transmission procedure into several segments to achieve optimal layer-wise communications and computations overlapping during run-time. Through experiments, we verify that DynaComm manages to achieve optimal layer-wise scheduling for all cases compared to competing strategies while the model accuracy remains untouched.

Mengdi Liu, Guangquan Xu

Cold Start (CS) and sparse evaluation problems dramatically degrade recommendation performance in large-scale recommendation systems such as Taobao and eBay. We name this degradation as the sparse trust problem, which will cause the decrease of the recommendation accuracy rate. To address this problem we propose a novel sparse trust mining method, which is based on the Roundtable Gossip Algorithm (RGA). First, we define the relevant representation of sparse trust, which provides a research idea to solve the problem of sparse evidence in the large-scale recommendation system. Based on which the RGA is proposed for mining latent sparse trust relationships between entities in large-scale recommendation systems. Second, we propose an efficient and simple anti-sparsification method, which overcomes the disadvantages of random trust relationship propagation and Grade Inflation caused by different users have different standard for item rating. Finally, the experimental results show that our method can effectively mine new trust relationships and mitigate the sparse trust problem.