Biblio

In big data applications such as healthcare data mining, due to privacy concerns, it is necessary to collect predictions from multiple information sources for the same instance, with raw features being discarded or withheld when aggregating multiple predictions. Besides, crowd-sourced labels need to be aggregated to estimate the ground truth of the data. Due to the imperfection caused by predictive models or human crowdsourcing workers, noisy and conflicting information is ubiquitous and inevitable. Although state-of-the-art aggregation methods have been proposed to handle label spaces with flat structures, as the label space is becoming more and more complicated, aggregation under a label hierarchical structure becomes necessary but has been largely ignored. These label hierarchies can be quite informative as they are usually created by domain experts to make sense of highly complex label correlations such as protein functionality interactions or disease relationships. We propose a novel multi-source hierarchical prediction consolidation method to effectively exploits the complicated hierarchical label structures to resolve the noisy and conflicting information that inherently originates from multiple imperfect sources. We formulate the problem as an optimization problem with a closed-form solution. The consolidation result is inferred in a totally unsupervised, iterative fashion. Experimental results on both synthetic and real-world data sets show the effectiveness of the proposed method over existing alternatives.

The Smart Grid control systems need to be protected from internal attacks within the perimeter. In Smart Grid, the Intelligent Electronic Devices (IEDs) are resource-constrained devices that do not have the ability to provide security analysis and protection by themselves. And the commonly used industrial control system protocols offer little security guarantee. To guarantee security inside the system, analysis and inspection of both internal network traffic and device status need to be placed close to IEDs to provide timely information to power grid operators. For that, we have designed a unique, extensible and efficient operation-level traffic analyzer framework. The timing evaluation of the analyzer overhead confirms efficiency under Smart Grid operational traffic.

In March 2016, several online news media reported on the inadequate emotional capabilities of interactive virtual assistants. While significant progress has been made in the general intelligence and functionality of virtual agents (VA), the emotional intelligent (EI) VA has yet been thoroughly explored. We examine user's perception of EI of virtual agents through Zara The Supergirl, a virtual agent that conducts question and answering type of conversational testing and counseling online. The results show that overall users perceive an emotion-expressing VA (EEVA) to be more EI than a non-emotion-expressing VA (NEEVA). However, simple affective expression may not be sufficient enough for EEVA to be perceived as fully EI.

We propose a multi party conversational social interface NAMIDA through a pilot study. The system consists of three robots that can converse with each other about environment throughout the road. Through this model, the directed utterances towards the driver diminishes by utilizing turn-taking process between the agents, and the mental workload of the driver can be reduced compared to the conventional one-to-one communication based approach that directly addresses the driver. We set up an experiment to compare the both approaches to explore their effects on the workload and attention behaviors of drivers. The results indicated that the multi-party conversational approach has a better effect on reducing certain workload factors. Also, the analysis of attention behaviors of drivers revealed that our method can better promote the drivers to focus on the road.

Heap buffer overflow (underflow) errors are a common source of security vulnerabilities. One prevention mechanism is to add object bounds meta information and to instrument the program with explicit bounds checks for all memory access. The so-called "fat pointers" approach is one method for maintaining and propagating the meta information where native machine pointers are replaced with "fat" objects that explicitly store object bounds. Another approach is "low fat pointers", which encodes meta information within a native pointer itself, eliminating space overheads and also code compatibility issues. This paper presents a new low-fat pointer encoding that is fully compatible with existing libraries (e.g. pre-compiled libraries unaware of the encoding) and standard hardware (e.g. x86\_64). We show that our approach has very low memory overhead, and competitive with existing state-of-the-art bounds instrumentation solutions.

Trust plays an important role in various user-facing systems and applications. It is particularly important in the context of decision support systems, where the system's output serves as one of the inputs for the users' decision making processes. In this work, we study the dynamics of explicit and implicit user trust in a simulated automated quality monitoring system, as a function of the system accuracy. We establish that users correctly perceive the accuracy of the system and adjust their trust accordingly.

Differential privacy enables organizations to collect accurate aggregates over sensitive data with strong, rigorous guarantees on individuals' privacy. Previous work has found that under differential privacy, computing multiple correlated aggregates as a batch, using an appropriate strategy, may yield higher accuracy than computing each of them independently. However, finding the best strategy that maximizes result accuracy is non-trivial, as it involves solving a complex constrained optimization program that appears to be non-convex. Hence, in the past much effort has been devoted in solving this non-convex optimization program. Existing approaches include various sophisticated heuristics and expensive numerical solutions. None of them, however, guarantees to find the optimal solution of this optimization problem. This paper points out that under (ε, ཬ)-differential privacy, the optimal solution of the above constrained optimization problem in search of a suitable strategy can be found, rather surprisingly, by solving a simple and elegant convex optimization program. Then, we propose an efficient algorithm based on Newton's method, which we prove to always converge to the optimal solution with linear global convergence rate and quadratic local convergence rate. Empirical evaluations demonstrate the accuracy and efficiency of the proposed solution.

Text-based Captchas have been widely used to deter misuse of services on the Internet. However, many designs have been broken. It is intellectually interesting and practically relevant to look for alternative designs, which are currently a topic of active research. We motivate the study of Chinese Captchas as an interesting alternative design - co-unterintuitively, it is possible to design Chinese Captchas that are universally usable, even to those who have never studied Chinese language. More importantly, we ask a fundamental question: is the segmentation-resistance principle established for Roman-character based Captchas applicable to Chinese based designs? With deep learning techniques, we offer the first evidence that computers do recognize individual Chinese characters well, regardless of distortion levels. This suggests that many real-world Chinese schemes are insecure, in contrast to common beliefs. Our result offers an essential guideline to the design of secure Chinese Captchas, and it is also applicable to Captchas using other large-alphabet languages such as Japanese.

Non-malleability is an important and intensively studied security notion for many cryptographic primitives. In the context of public key encryption, this notion means it is infeasible for an adversary to transform an encryption of some message m into one of a related message m' under the given public key. Although it has provided a strong security property for many applications, it still does not suffice for some scenarios like the system where the users could issue keys on-the-fly. In such settings, the adversary may have the power to transform the given public key and the ciphertext. To withstand such attacks, Fischlin introduced a stronger notion, known as complete non-malleability, which requires that the non-malleability property be preserved even for the adversaries attempting to produce a ciphertext of some related message under the transformed public key. To date, many schemes satisfying this stronger security have been proposed, but they are either inefficient or proved secure in the random oracle model. In this work, we put forward a new encryption scheme in the common reference string model. Based on the standard DBDH assumption, the proposed scheme is proved completely non-malleable secure against adaptive chosen ciphertext attacks in the standard model. In our scheme, the well-formed public keys and ciphertexts could be publicly recognized without drawing support from unwieldy techniques like non-interactive zero knowledge proofs or one-time signatures, thus achieving a better performance.

A finite state machine (FSM) is responsible for controlling the overall functionality of most digital systems and, therefore, the security of the whole system can be compromised if there are vulnerabilities in the FSM. These vulnerabilities can be created by improper designs or by the synthesis tool which introduces additional don't-care states and transitions during the optimization and synthesis process. An attacker can utilize these vulnerabilities to perform fault injection attacks or insert malicious hardware modifications (Trojan) to gain unauthorized access to some specific states. To our knowledge, no systematic approaches have been proposed to analyze these vulnerabilities in FSM. In this paper, we develop a framework named Analyzing Vulnerabilities in FSM (AVFSM) which extracts the state transition graph (including the don't-care states and transitions) from a gate-level netlist using a novel Automatic Test Pattern Generation (ATPG) based approach and quantifies the vulnerabilities of the design to fault injection and hardware Trojan insertion. We demonstrate the applicability of the AVFSM framework by analyzing the vulnerabilities in the FSM of AES and RSA encryption module. We also propose a low-cost mitigation technique to make FSM more secure against these attacks.

Real world wireless networks usually have diverse connectivity characteristics. Although existing works have identified replication as the key to the successful design of routing protocols for these networks, the questions of when the replication should be used, by how much, and how to distribute packet copies are still not satisfactorily answered. In this paper, we investigate the above questions and present the design of the Hybrid Routing Protocol (HRP). We make a key observation that delay correlations can significantly impact performance improvements gained from packet replication. Thus, we propose a novel model to capture the correlations of inter-contact times among a group of nodes. HRP utilizes both direct delays feedback and the proposed model to estimate the replication gain, which is then fed into a novel regret-minimization algorithm to dynamically decide the amount of packet replication under unknown network conditions. We evaluate HRP through extensive simulations. We show that HRP achieves up to 3.5x delivery ratio improvement and up to 50% delay reduction, with comparable and even lower overhead than state-of-art routing protocols.

Multiple-input multiple-output (MIMO) techniques have been the subject of increased attention for underwater acoustic communication for its ability to significantly improve the channel capabilities. Recently, an under-ice MIMO acoustic communication experiment was conducted in shallow water which differs from previous works in that the water column was covered by about 40 centimeters thick sea ice. In this experiment, high frequency MIMO signals centered at 10 kHz were transmitted from a two-element source array to a four-element vertical receive array at 1km range. The unique under-ice acoustic propagation environment in shallow water seems naturally separate data streams from different transducers, but there is still co-channel interference. Time reversal followed by a single channel decision feedback equalizer is used in this paper to compensate for the inter-symbol interference and co-channel interference. It is demonstrated that this simple receiver scheme is good enough to realize robust performance using fewer hydrophones (i.e. 2) without the explicit use of complex co-channel interference cancelation algorithms such as parallel interference cancelation or serial interference cancelation. Two channel estimation algorithms based on least square and least mean square are also studied for MIMO communications in this paper and their performance are compared using experimental data.

The local minima has been deemed as a challenging issue when designing routing protocols for 3D underwater sensor networks. Recently, harmonic potential field method has been used to tackle the issue of local minima which was also a major bottleneck in path planning and obstacle avoidance of robotics community. Inspired by this, this paper proposes a harmonic potential field based routing protocol for 3D underwater sensor networks with local minima. More specifically, the harmonic potential field is calculated using harmonic functions and Dirichlet boundary conditions are used for the local minima, sink(or seabuoy) and sending node. Numerical results show the effectiveness of the proposed routing protocol.

The harsh environment in the water has imposed challenges for underwater sensor networks (USNs), which collect the sensed data from the underwater sensors to the sink on land. The time-varying underwater acoustic channel has low band-width and high bit error rate, which leads to low data collection efficiency. Furthermore, the heterogeneous model of USNs that uses acoustic communications under the water and wireless communication above the water makes it difficult in efficient routing and forwarding for data collection. To this end, we propose a novel on-surface wireless-assisted opportunistic routing (SurOpp) for USNs. SurOpp deploys multiple buoy nodes on surface and includes all of them in the forwarding candidates to form a receive diversity. The opportunities of reception and forwarding in buoy nodes are exploited to improve the end-to-end transmissions. SurOpp also adopts rateless codes in the source to achieve opportunistic reception in the sink. The cooperation of both opportunistic reception in the buoys and the sink further decreases the messages of control overhead. The wireless interface in the buoy undertakes all the message exchanges in forwarding coordination to compensate the bandwidth limit of the acoustic channel. Simulations in NS3 show that SurOpp outperforms the traditional routing and existing opportunistic routing in terms of packet delivery ratio, end-to-end delay and energy consumption.

Encryption is critical in protecting the confidentiality of users' data on mobile devices. However, research has shown that many mobile apps are not correctly using the ciphers, which makes them vulnerable to the attacks. The existing resources on cipher programming education do not provide enough practical scenarios to help students learn the cipher programming in the context of real world situations with programs that have complex interacting modules with access to networking, storage, and database. This poster introduces a course module that teaches students how to develop secure Android applications by correctly using Android's cryptography APIs. This course module is targeted to two areas where programmers commonly make many mistakes: password based encryption and SSL certificate validation. The core of the module includes a real world sample Android program for students to secure by implementing cryptographic components correctly. The course module will use open-ended problem solving to let students freely explore the multiple options in securing the application. The course module includes a lecture slide on Android's Crypto library, its common misuses, and suggested good practices. Assessment materials will also be included in the course module. This course module will be used and evaluated in a Network Security class. We will present the results of the evaluation in the conference.

In this paper, we present a novel acoustic sensing technique that recognizes two convenient input actions: hand gestures and on-body touch. We achieved them by observing the frequency spectrum of the wave propagated in the body, around the periphery of the wrist. Our approach can recognize hand gestures and on-body touch concurrently in real-time and is expected to obtain rich input variations by combining them. We conducted a user study that showed classification accuracy of 97%, 96%, and 97% for hand gestures, touches on the forearm, and touches on the back of the hand.

We consider 3-dimensional(3D) underwater random network (UAN) where the nodes are uniformly distributed in a cuboid region. Then we derive the closed-form probability of finding a receiver in an ellipsoid neighborhood of an arbitrary sender. Computer simulation shows that the analytical result is generally consistent with the simulated result.

The popularity of cloud hosting services also brings in new security challenges: it has been reported that these services are increasingly utilized by miscreants for their malicious online activities. Mitigating this emerging threat, posed by such "bad repositories" (simply Bar), is challenging due to the different hosting strategy to traditional hosting service, the lack of direct observations of the repositories by those outside the cloud, the reluctance of the cloud provider to scan its customers' repositories without their consent, and the unique evasion strategies employed by the adversary. In this paper, we took the first step toward understanding and detecting this emerging threat. Using a small set of "seeds" (i.e., confirmed Bars), we identified a set of collective features from the websites they serve (e.g., attempts to hide Bars), which uniquely characterize the Bars. These features were utilized to build a scanner that detected over 600 Bars on leading cloud platforms like Amazon, Google, and 150K sites, including popular ones like groupon.com, using them. Highlights of our study include the pivotal roles played by these repositories on malicious infrastructures and other important discoveries include how the adversary exploited legitimate cloud repositories and why the adversary uses Bars in the first place that has never been reported. These findings bring such malicious services to the spotlight and contribute to a better understanding and ultimately eliminating this new threat.

Simulation and verification of the on-die power delivery network (PDN) is one of the key steps in the design of integrated circuits (ICs). With the very large sizes of modern grids, verification of PDNs has become very expensive and a host of techniques for faster simulation and grid model approximation have been proposed. These include topological node elimination, as in TICER and full-blown numerical model order reduction (MOR) as in PRIMA and related methods. However, both of these traditional approaches suffer from certain drawbacks that make them expensive and limit their scalability to very large grids. In this paper, we propose a novel technique for grid reduction that is a hybrid of both approaches–-the method is numerical but also factors in grid topology. It works by eliminating whole internal layers of the grid at a time, while aiming to preserve the dynamic behavior of the resulting reduced grid. Effectively, instead of traditional node-by-node topological elimination we provide a numerical layer-by-layer block-matrix approach that is both fast and accurate. Experimental results show that this technique is capable of handling very large power grids and provides a 4.25x speed-up in transient analysis.

Since debugging is a time-consuming activity, automated program repair tools such as GenProg have garnered interest. A recent study revealed that the majority of GenProg repairs avoid bugs simply by deleting functionality. We found that SPR, a state-of-the-art repair tool proposed in 2015, still deletes functionality in their many "plausible" repairs. Unlike generate-and-validate systems such as GenProg and SPR, semantic analysis based repair techniques synthesize a repair based on semantic information of the program. While such semantics-based repair methods show promise in terms of quality of generated repairs, their scalability has been a concern so far. In this paper, we present Angelix, a novel semantics-based repair method that scales up to programs of similar size as are handled by search-based repair tools such as GenProg and SPR. This shows that Angelix is more scalable than previously proposed semantics based repair methods such as SemFix and DirectFix. Furthermore, our repair method can repair multiple buggy locations that are dependent on each other. Such repairs are hard to achieve using SPR and GenProg. In our experiments, Angelix generated repairs from large-scale real-world software such as wireshark and php, and these generated repairs include multi-location repairs. We also report our experience in automatically repairing the well-known Heartbleed vulnerability.

The malicious JavaScript is a common springboard for attackers to launch several types of network attacks, such as Drive-by-Download and malicious PDF delivery attack. In order to elude detection of signature matching, malicious JavaScript is often packed (so-called "obfuscation") with diversified algorithms therefore the occurrence of obfuscation is always a good pointer for potential maliciousness. In this investigation, we propose a light weight approach for quickly filtering obfuscated JavaScript by a novel method of tokenizing JavaScript text at letter level and information-theoretic measures, based on the previous work in the domain of detecting obfuscated malicious code as well as the pattern analysis of natural languages. The new approach is apparently time efficient compared to existing systems since it processes much less objects while it is also proved to be able to reach the acceptable detection accuracies.

Automatic face recognition techniques applied on particular group or mass database introduces error cases. Error prevention is crucial for the court. Reranking of recognition results based on anthropology analysis can significant improve the accuracy of automatic methods. Previous studies focused on manual facial comparison. This paper proposed a weighted facial similarity computing method based on morphological analysis of components characteristics. Search sequence of face recognition reranked according to similarity, while the interference terms can be removed. Within this research project, standardized photographs, surveillance videos, 3D face images, identity card photographs of 241 male subjects from China were acquired. Sequencing results were modified by modeling selected individual features from the DMV altas. The improved method raises the accuracy of face recognition through anthroposophic or morphologic theory.

Heterogeneous face recognition aims to identify or verify person identity by matching facial images of different modalities. In practice, it is known that its performance is highly influenced by modality inconsistency, appearance occlusions, illumination variations and expressions. In this paper, a new method named as ensemble of sparse cross-modal metrics is proposed for tackling these challenging issues. In particular, a weak sparse cross-modal metric learning method is firstly developed to measure distances between samples of two modalities. It learns to adjust rank-one cross-modal metrics to satisfy two sets of triplet based cross-modal distance constraints in a compact form. Meanwhile, a group based feature selection is performed to enforce that features in the same position of two modalities are selected simultaneously. By neglecting features that attribute to "noise" in the face regions (eye glasses, expressions and so on), the performance of learned weak metrics can be markedly improved. Finally, an ensemble framework is incorporated to combine the results of differently learned sparse metrics into a strong one. Extensive experiments on various face datasets demonstrate the benefit of such feature selection especially when heavy occlusions exist. The proposed ensemble metric learning has been shown superiority over several state-of-the-art methods in heterogeneous face recognition.

In the past three years, Emotion Recognition in the Wild (EmotiW) Grand Challenge has drawn more and more attention due to its huge potential applications. In the fourth challenge, aimed at the task of video based emotion recognition, we propose a multi-clue emotion fusion (MCEF) framework by modeling human emotion from three mutually complementary sources, facial appearance texture, facial action, and audio. To extract high-level emotion features from sequential face images, we employ a CNN-RNN architecture, where face image from each frame is first fed into the fine-tuned VGG-Face network to extract face feature, and then the features of all frames are sequentially traversed in a bidirectional RNN so as to capture dynamic changes of facial textures. To attain more accurate facial actions, a facial landmark trajectory model is proposed to explicitly learn emotion variations of facial components. Further, audio signals are also modeled in a CNN framework by extracting low-level energy features from segmented audio clips and then stacking them as an image-like map. Finally, we fuse the results generated from three clues to boost the performance of emotion recognition. Our proposed MCEF achieves an overall accuracy of 56.66% with a large improvement of 16.19% with respect to the baseline.

In this paper, we describe the formatting guidelines for ACM SIG Proceedings. In order to assure safety of Chinese Train Control System (CTCS), it is necessary to ensure the operational risk is acceptable throughout its life-cycle, which requires a pragmatic risk assessment required for effective risk control. Many risk assessment techniques currently used in railway domain are qualitative, and rely on the experience of experts, which unavoidably brings in subjective judgements. This paper presents a method that combines fuzzy reasoning and analytic hierarchy process approach to quantify the experiences of experts to get the scores of risk parameters. Fuzzy reasoning is used to obtain the risk of system hazard, analytic hierarchy process approach is used to determine the risk level (RL) and its membership of the system. This method helps safety analyst to calculate overall collective risk level of system. A case study of risk assessment of CTCS system is used to demonstrate this method can give quantitative result of collective risks without much information from experts, but can support the risk assessment with risk level and its membership, which are more valuable to guide the further risk management.