Streaming variational inference for dirichlet process mixtures

Bayesian nonparametric models are theoretically suitable to learn streaming data due to their complexity relaxation to the volume of observed data. However, most of the existing variational inference algorithms are not applicable to streaming applications since they re-quire truncation on variational distributions. In this paper, we present two truncation-free variational algorithms, one for mix-membership inference called TFVB (truncation-free variational Bayes), and the other for hard clustering inference called TFME (truncation-free maximization expectation). With these algorithms, we further developed a streaming learning framework for the popular Dirichlet process mixture (DPM) models. Our ex-periments demonstrate the usefulness of our framework in both synthetic and real-world data.

Maximizing real-time streaming services based on a multi-servers networking framework

In recent years, we have witnessed substantial exploitation of real-time streaming applications, such as video surveillance system on road crosses of a city. So far, real world applications mainly rely on the traditional well-known client-server and peer-to-peer schemes as the fundamental mechanism for communication. However, due to the limited resources on each terminal device in the applications, these two schemes cannot well leverage the processing capability between the source and destination of the video traffic, which leads to limited streaming services. For this reason, many QoS sensitive application cannot be supported in the real world. In this paper, we are motivated to address this problem by proposing a novel multi-server based framework. In this framework, multiple servers collaborate with each other to form a virtual server (also called cloud-server), and provide high-quality services such as real-time streams delivery and storage. Based on this framework, we further introduce a (1-?) approximation algorithm to solve the NP-complete "maximum services"(MS) problem with the intention of handling large number of streaming flows originated by networks and maximizing the total number of services. Moreover, in order to backup the streaming data for later retrieval, based on the framework, an algorithm is proposed to implement backups and maximize streaming flows simultaneously. We conduct a series of experiments based on simulations to evaluate the performance of the newly proposed framework. We also compare our scheme to several traditional solutions. The results suggest that our proposed scheme significantly outperforms the traditional solutions.

A QoE centric distributed caching approach for vehicular video streaming in cellular networks

Distributed caching-empowered wireless networks can greatly improve the efficiency of data storage and transmission and thereby the users' quality of experience (QoE). However, how this technology can alleviate the network access pressure while ensuring the consistency of content delivery is still an open question, especially in the case where the users are in fast motion. Therefore, in this paper, we investigate the caching issue emerging from a forthcoming scenario where vehicular video streaming is performed under cellular networks. Specifically, a QoE centric distributed caching approach is proposed to fulfill as many users' requests as possible, considering the limited caching space of base stations and basic user experience guarantee. Firstly, a QoE evaluation model is established using verified empirical data. Also, the mathematic relationship between the streaming bit rate and actual storage space is developed. Then, the distributed caching management for vehicular video streaming is formulated as a constrained optimization problem and solved with the generalized-reduced gradient method. Simulation results indicate that our approach can improve the users' satisfaction ratio by up to 40%.

A performance evaluation of machine learning-based streaming spam tweets detection

The popularity of Twitter attracts more and more spammers. Spammers send unwanted tweets to Twitter users to promote websites or services, which are harmful to normal users. In order to stop spammers, researchers have proposed a number of mechanisms. The focus of recent works is on the application of machine learning techniques into Twitter spam detection. However, tweets are retrieved in a streaming way, and Twitter provides the Streaming API for developers and researchers to access public tweets in real time. There lacks a performance evaluation of existing machine learning-based streaming spam detection methods. In this paper, we bridged the gap by carrying out a performance evaluation, which was from three different aspects of data, feature, and model. A big ground-truth of over 600 million public tweets was created by using a commercial URL-based security tool. For real-time spam detection, we further extracted 12 lightweight features for tweet representation. Spam detection was then transformed to a binary classification problem in the feature space and can be solved by conventional machine learning algorithms. We evaluated the impact of different factors to the spam detection performance, which included spam to nonspam ratio, feature discretization, training data size, data sampling, time-related data, and machine learning algorithms. The results show the streaming spam tweet detection is still a big challenge and a robust detection technique should take into account the three aspects of data, feature, and model.

Streaming analysis in wireless sensor networks

Two new incremental models for online anomaly detection in data streams at nodes in wireless sensor networks are discussed. These models are incremental versions of a model that uses ellipsoids to detect first, second, and higher-ordered anomalies in arrears. The incremental versions can also be used this way but have additional capabilities offered by processing data incrementally as they arrive in time. Specifically, they can detect anomalies 'on-the-fly' in near real time. They can also be used to track temporal changes in near real-time because of sensor drift, cyclic variation, or seasonal changes. One of the new models has a mechanism that enables graceful degradation of inputs in the distant past (fading memory). Three real datasets from single sensors in deployed environmental monitoring networks are used to illustrate various facets of the new models. Examples compare the incremental version with the previous batch and dynamic models and show that the incremental versions can detect various types of dynamic anomalies in near real time.

Generalized pseudo-likelihood estimates for markov random fields on lattice

In this paper we generalize Besag's pseudo-likelihood function for spatial statistical models on a region of a lattice. The correspondingly defined maximum generalized pseudo-likelihood estimates (MGPLEs) are natural extensions of Besag's maximum pseudo-likelihood estimate (MPLE). The MGPLEs connect the MPLE and the maximum likelihood estimate. We carry out experimental calculations of the MGPLEs for spatial processes on the lattice. These simulation results clearly show better performances of the MGPLEs than the MPLE, and the performances of differently defined MGPLEs are compared. These are also illustrated by the application to two real data sets.

Modeling dialogue with mixed initiative in design space exploration

Exploration with a generative formalism must necessarily account for the nature of interaction between humans and the design space explorer. Established accounts of design interaction are made complicated by two propositions in Woodbury and Burrow's Keynote on design space exploration. First, the emphasis on the primacy of the design space as an ordered collection of partial designs (version, alternatives, extensions). Few studies exist in the design interaction literature on working with multiple threads simultaneously. Second, the need to situate, aid, and amplify human design intentions using computational tools. Although specific research and practice tools on amplification (sketching, generation, variation) have had success, there is a lack of generic, flexible, interoperable, and extensible representation to support amplification. This paper addresses the above, working with design threads and computer-assisted design amplification through a theoretical model of dialogue based on Grice's model of rational conversation. Using the concept of mixed initiative, the paper presents a visual notation for representing dialogue between designer and design space formalism through abstract examples of exploration tasks and dialogue integration.

Frequency distribution of TATA box and extension sequences on human promoters

TATA box is one of the most important transcription factor binding sites. But the exact sequences of TATA box are still not very clear yet. In this study, we conducted a dedicated analysis on the frequency distribution of TATA Box and its extension sequences on human promoters. Sixteen TATA elements derived from TATA Box motif, TATAWAWN, were classified into three distribution patterns: peak, bottom-peak and bottom. Fourteen TATA extension sequences (up to two base extensions) were predicted to be the new TATA Box elements because of their high motif factors, which indicate their statistical significance. Statistical analysis on the promoters of mouse, zebrafish and drosophila melanogaster verified seven of these elements. It was also observed that the distribution of TATA elements on the promoters of housekeeping genes are very similar with their distribution on the promoters of tissue specific genes in human.

Defining the scope of practice of the emergency nurse practitioner role in a metropolitan emergency department

This study examined the emergency nurse practitioner candidate (ENPC) scope of practice in a Victorian emergency department (ED). The emergency nurse practitioner (ENP) role is relatively new in Victoria and the scope of the ENP(C) practice is yet to be defined. International research literature regarding the ENP role has focused on outcomes such as patient satisfaction, waiting times and/or ED length of stay, accuracy and adequacy of documentation, use of radiography, and patient education, health promotion and communication issues. A prospective exploratory design was used to conduct this cohort study. There were 476 ENPC-managed patients between 14 July 2004 and 31 March 2005 with an average age of 29 years. The majority (77.2%) of ENPC-managed patients were discharged from the ED. The majority of the ENPC time was devoted to clinical practice (55%) and development of clinical practice guidelines (25%). Of patients managed by the ENPC, 49.6% required medications, 51% required diagnostic imaging and 8.6% required pathology testing during their ED stay. The most common discharge referrals were made to local medical officers (73.5%) and the most common referrals made for patients requiring admission were made to the plastic surgery (37.3%) and orthopaedic (35.5%) units. Extensions to the current scope of emergency nursing practice are pivotal to effective management of specific patient groups by ENP. The ENP model of care is an important strategy for the management of increased service demands in Victoria; however, little is known about the scope of the ENPC practice and many outcomes of the ENP care are yet to be defined. Further research to better understand the relationships between ENP outcomes is required if the contribution that ENPs make to emergency care is to be accurately quantified.