In this paper, we propose a novel Heterogeneous Gaussian Mechanism (HGM) to preserve differential privacy in deep neural networks, with provable robustness against adversarial examples. We first relax the constraint of the privacy budget in the traditional Gaussian Mechanism from (0, 1] to (0, ∞), with a new bound of the noise scale to preserve differential privacy. The noise in our mechanism can be arbitrarily redistributed, offering a distinctive ability to address the trade-off between model utility and privacy loss. To derive provable robustness, our HGM is applied to inject Gaussian noise into the first hidden layer. Then, a tighter robustness bound is proposed. Theoretical analysis and thorough evaluations show that our mechanism notably improves the robustness of differentially private deep neural networks, compared with baseline approaches, under a variety of model attacks.

Abuse of prescription drugs and of illicit drugs has been declared a "national emergency" [1]. This crisis includes the misuse and abuse of cannabinoids, opioids, tranquilizers, stimulants, inhalants, and other types of psychoactive drugs, which statistical analysis documents as a rising trend in the United States. The most recent reports from the National Survey on Drug Use and Health (NSDUH) [2] estimate that 10.6% of the total population of people ages 12 years and older (i.e., about 28.6 million people) misused illicit drugs in 2016, which represents an increase of 0.5% since 2015 [3]. According to the Centers for Disease Control and Prevention (CDC), opioid drugs were involved in 42,249 known deaths in 2016 nationwide [4]. In addition, the number of heroin-involved deaths has been increasing sharply for 5 years, and surpassed the number of firearm homicides in 2015 [5]. In April 2017, the Department of Health and Human Services announced their "Opioid Strategy" to battle the country's drug-abuse crisis [1]. In the Opioid Strategy, one of the major aims is to strengthen public health data collection, to inform a timeliness

No abstract

International audienceRecent improvements in positioning technology have led to a much wider availability of massive moving object data. A crucial task is to find the moving objects that travel together. In common, these object sets are called object movement patterns. Due to the emergence of many different kinds of object movement patterns in recent years, different approaches have been proposed to extract them. However, each approach only focuses on mining a specific kind of patterns. It is costly and time consuming to mine and manage various number of patterns, since we have to execute a large number of different pattern mining algorithms. Moreover, we have to execute these algorithms again whenever new data are added to the existing database. To address these issues, we first redefine movement patterns in the itemset context. Second, we propose a unifying approach, named GeT_Move, which uses a frequent closed itemset-based object movement pattern-mining algorithm to mine and manage different patterns. GeT_Move is developed in two versions which are GeT_Move and Incremental GeT_Move. To optimize the efficiency and to free the parameters setting, we further propose a Parameter Free Incremental GeT_Move algorithm. Comprehensive experiments are performed on real and large synthetic datasets to demonstrate the effectiveness and efficiency of our approaches

International audienceGradual patterns highlight covariations of attributes of the form " The more/less X, the more/less Y ". Their usefulness in several applications has recently stimulated the synthesis of several algorithms for their automated discovery from large datasets. However, existing techniques require all the interesting data to be in a single database relation or table. This paper extends the notion of gradual pattern to the case in which the co-variations are possibly expressed between attributes of different database relations. The interestingness measure for this class of " relational gradual patterns " is defined on the basis of both Kendall's τ and gradual supports. Moreover, this paper proposes two algorithms, named τ RGP Miner and gRGP Miner, for the discovery of relational gradual rules. Three pruning strategies to reduce the search space are proposed. The efficiency of the algorithms is empirically validated, and the usefulness of relational gradual patterns is proved on some real-world databases

Finding similar users is one of the probable applications in social media. The similarity between users can be measured in two different approaches: the semantic similarity and the similarity in terms of social relations. These two approaches can be combined with different weight factors. However, the conventional combination scheme has a critical drawback that the weight factors are fixed for every user and thus it is not optimized at those users that are using rare terms or do not have sufficient relations with other users. To address this problem, in this paper, we propose an adaptive combination scheme of tag-based similarity and link-based similarity in which the weight factors are dynamically determined for each user by evaluating each user's characteristics such as tag commonness and link strength. The experimental results with a Flickr data set show that the proposed scheme consistently outperforms the previous work by about 20%.

In this work, we study a basic and practically important strategy to help prevent and/or delay an outbreak in the context of network: limiting the contact between individuals. In this paper, we introduce the average neighborhood size as a new measure for the degree of being small-world and utilize it to formally define the desmall-world network problem. We also prove the NP-hardness of the general reachable pair cut problem and propose a greedy edge betweenness based approach as the benchmark in selecting the candidate edges for solving our problem. Furthermore, we transform the de-small-world network problem as an OR-AND Boolean function maximization problem, which is also an NP-hardness problem. In addition, we develop a numerical relaxation approach to solve the Boolean function maximization and the de-small-world problem. Also, we introduce the short-betweenness, which measures the edge importance in terms of all short paths with distance no greater than a certain threshold, and utilize it to speed up our numerical relaxation approach. The experimental evaluation demonstrates the effectiveness and efficiency of our approaches. underlying (social) network less small-world, or simply "de-smallworld", i.e., the distances between individuals increase to delay the spreading process. In many situations, such a strategy is often easily and even voluntarily adopted. For instance, during the SARS epidemic in Beijing, 2004, there are much less people appearing in the public places. In addition, this approach can also be deployed in complement to the quarantine approach.