Mengjun Xie

Mining Twitter to Assess the Public Perception of the "Internet of Things".

Social media analysis has shown tremendous potential to understand publics opinion on a wide variety of topics. In this paper, we have mined Twitter to understand the publics perception of the Internet of Things (IoT). We first generated the discussion trends of the IoT from multiple Twitter data sources and validated these trends with Google Trends. We then performed sentiment analysis to gain insights of the public’s attitude towards the IoT. As anticipated, our analysis indicates that the publics perception of the IoT is predominantly positive. Further, through topic modeling, we learned that public tweets discussing the IoT were often focused on business and technology. However, the public has great concerns about privacy and security issues toward the IoT based on the frequent appearance of related terms. Nevertheless, no unexpected perceptions were identified through our analysis. Our analysis was challenged by the limited fraction of tweets relevant to our study. Also, the user demographics of Twitter users may not be strongly representative of the population of the general public.

CollaborationViz: interactive visual exploration of biomedical research collaboration networks.

Social network analysis (SNA) helps us understand patterns of interaction between social entities. A number of SNA studies have shed light on the characteristics of research collaboration networks (RCNs). Especially, in the Clinical Translational Science Award (CTSA) community, SNA provides us a set of effective tools to quantitatively assess research collaborations and the impact of CTSA. However, descriptive network statistics are difficult for non-experts to understand. In this article, we present our experiences of building meaningful network visualizations to facilitate a series of visual analysis tasks. The basis of our design is multidimensional, visual aggregation of network dynamics. The resulting visualizations can help uncover hidden structures in the networks, elicit new observations of the network dynamics, compare different investigators and investigator groups, determine critical factors to the network evolution, and help direct further analyses. We applied our visualization techniques to explore the biomedical RCNs at the University of Arkansas for Medical Sciences – a CTSA institution. And, we created CollaborationViz, an open-source visual analytical tool to help network researchers and administration apprehend the network dynamics of research collaborations through interactive visualization.

CLARA: an integrated clinical research administration system

Administration of human subject research is complex, involving not only the institutional review board but also many other regulatory and compliance entities within a research enterprise. Its efficiency has a direct and substantial impact on the conduct and management of clinical research. In this paper, we report on the Clinical Research Administration (CLARA) platform developed at the University of Arkansas for Medical Sciences. CLARA is a comprehensive web-based system that can streamline research administrative tasks such as submissions, reviews, and approval processes for both investigators and different review committees on a single integrated platform. CLARA not only helps investigators to meet regulatory requirements but also provides tools for managing other clinical research activities including budgeting, contracting, and participant schedule planning.

Comparison of PIN- and pattern-based behavioral biometric authentication on mobile devices

Personal identification numbers (PIN) and unlock patterns are highly popular authentication mechanisms on smart mobile devices but they are not sufficiently secure. PIN or pattern mechanisms enhanced by additional, implicit behavioral biometric authentication can offer stronger authentication assurance while preserving usability, therefore becoming very attractive. Individual studies on PIN- and pattern-based behavioral biometric authentication on smartphones were conducted but their results cannot be directly compared. In this work, we present a comparison study on the authentication accuracy between PIN-based and pattern-based behavioral biometric authentication using both smartphone and tablet. We developed a uniform framework for both PIN-based and pattern-based schemes and used two representative methods-Histogram and DTW-for user verification. We recruited 15 users and collected behavioral biometric data for both simple and complex PINs and patterns. Our experimental results show that PIN-based and pattern-based behavioral biometric authentication schemes can achieve about the same level of accuracy but not all verification methods are equal. The Histogram method can achieve more consistent results and handle template aging better than the DTW method based on our results. Our findings are expected to shed light on the exploration and analysis of effective behavioral biometric verification methods and facilitate more comprehensive investigation on behavioral biometric authentication for mobile devices.

LightGraph: Lighten Communication in Distributed Graph-Parallel Processing

A number of graph-structured computing abstractions have been proposed to address the needs of solving complex and large-scale graph algorithms. Distributed Graphlab and its successor, PowerGraph, are two such frameworks that have demonstrated excellent performance with high scalability and fault tolerance. However, excessive communication and state sharing among nodes in these frameworks not only reduce network efficiency but may also cause a decrease in runtime performance. In this paper, we first propose a mechanism that identifies and eliminates the avoidable communication during synchronization in existing distributed graph structured computing abstractions. We have implemented our method on PowerGraph and created LightGraph to reduce communication overhead in distributed graph-parallel computation systems. Furthermore, to minimize the required intra-graph synchronizations for PageRank-like applications, LightGraph also employs an edge direction-aware graph partitioning strategy, which optimally isolates the outgoing edges from the incoming edges of a vertex when creating and distributing replicas among different machines. We have conducted extensive experiments using real-world data, and our results verified the effectiveness of LightGraph. For example, when compared with the best existing graph placement method in PowerGraph, LightGraph can not only reduce up to 27.6% of synchronizing communication overhead for intra-graph synchronizations but also cut up to 17.1% runtime for PageRank.

Pre-execution Data Prefetching with Inter-thread I/O Scheduling

With the rate of computing power growing much faster than that of storage I/O access, parallel applications suffer more from I/O latency. I/O prefetching is effective in hiding I/O latency. However, existing I/O prefetching techniques are conservative and their effectiveness is limited. Recently, a more aggressive prefetching approach named pre-execution prefetching [19] has been proposed. In this paper, we first identify the drawback of this pre-execution prefetching approach, and then propose a new method to overcome the drawback by scheduling the I/O operations between the main thread and the prefetching thread. By careful I/O scheduling, our approach further extends the computation and I/O concurrency and avoids the I/O competition within one process. The results of extensive experiments, including experiments on real-life applications such as big matrix manipulation and Hill encryption, demonstrate the benefits of the proposed approach.

USign — A security enhanced electronic consent model

Electronic consent becomes increasingly popular in the healthcare sector given the many benefits it provides. However, security concerns, e.g., how to verify the identity of a person who is remotely accessing the electronic consent system in a secure and user-friendly manner, also arise along with the popularity of electronic consent. Unfortunately, existing electronic consent systems do not pay sufficient attention to those issues. They mainly rely on conventional password based authentication to verify the identity of an electronic consent user, which is far from being sufficient given that identity theft threat is real and significant in reality. In this paper, we present a security enhanced electronic consent model called USign. USign enhances the identity protection and authentication for electronic consent systems by leveraging handwritten signatures everyone is familiar with and mobile computing technologies that are becoming ubiquitous. We developed a prototype of USign and conducted preliminary evaluation on accuracy and usability of signature verification. Our experimental results show the feasibility of the proposed model.

Pre-execution data prefetching with I/O scheduling

Parallel applications suffer from I/O latency. Pre-execution I/O prefetching is effective in hiding I/O latency, in which a pre-execution prefetching thread is created and dedicated to fetch the data for the main thread in advance. However, existing pre-execution prefetching works do not pay attention to the relationship between the main thread and the pre-execution prefetching thread. They just simply pre-execute the I/O accesses using the prefetching thread as soon as possible failing to carefully coordinate them with the operations of the main thread. This drawback induces a series of adverse effects on pre-execution prefetching such as diminishing the degree of the parallelism between computation and I/O, delaying the I/O access of main threads, and aggravating the I/O resource competition in the whole system. In this paper, we propose a new method to overcome this drawback by scheduling the I/O operations among the main threads and the pre-execution prefetching threads. The results of extensive experiments on four popular benchmarks in parallel I/O performance area demonstrate the benefits of the proposed approach.

CamAuth: Securing Web Authentication with Camera

Frequent outbreak of password database leaks and server breaches in recent years manifests the aggravated security problems of web authentication using only password. Two-factor authentication, despite being more secure and strongly promoted, has not been widely applied to web authentication. Leveraging the unprecedented popularity of both personal mobile devices (e.g., Smartphones) and barcode scans through camera, we explore a new horizon in the design space of two-factor authentication. In this paper, we present CamAuth, a web authentication scheme that exploits pervasive mobile devices and digital cameras to counter various password attacks including man-in-the-middle and phishing attacks. In CamAuth, a mobile device is used as the second authentication factor to vouch for the identity of a use who is performing a web login from a PC. The device communicates directly with the PC through the secure visible light communication channels, which incurs no cellular cost and is immune to radio frequency attacks. CamAuth employs public-key cryptography to ensure the security of authentication process. We implemented a prototype system of CamAuth that consists of an Android application, a Chrome browser extension, and a Java-based web server. Our evaluation results indicate that CamAuth is a viable scheme for enhancing the security of web authentication.

A methodology for empirical analysis of brain connectivity through graph mining

Graph theoretical analysis has been applied to both structural and functional brain connectivity networks and has helped researchers conceive the effects of neurological and neuropsychiatric diseases including Alzhemier and Schizophrenia. However, existing graph theoretical approaches to brain connectivity networks simply assume that temporal correlations between brain regions are stable during the entire timeseries under consideration, and only focus on high-level network topological characteristics such as degree distribution. To advance the understanding of brain connectivity networks at a fine granularity, we propose a new method that can help discover connectivity-oriented insights from a time series of brain connectivity networks. In particular, our method is capable of identifying (1) strong correlations, which are represented as frequent edges in brain connectivity networks, for each individual subject, and (2) frequent substructures, which are connected components appearing frequently in brain connectivity networks, for a group of subjects. We apply the method to a data set of 38 subjects that were involved in a study of early life stress on depression development. Our findings have been echoed by the domain experts in terms of their clinical implications.