Thesis (Ph.D.) - University of St Andrews, April 2010.

Assignment (MComm)--University of Stellenbosch, 2005. / Bibliography. Also available via the Internet.

It has been long hypothesized that due to the inherent ambiguities of visual input and the limitations of the visual system, vision is a form of “unconscious inference” whereby the brain relies on assumptions (aka expectations) to interpret the external world. This hypothesis has been recently formalized into Bayesian models of perception (the “Bayesian brain”) that represent these expectations as prior probabilities. In this thesis, I focus on a particular kind of expectation that humans are thought to possess – that objects in the world tend to be still or move slowly – known as the “slow speed prior”. Through a combination of experimental and theoretical work, I investigate how the speed prior is acquired and how it impacts motion perception. The first part of my work consists of an experiment where subjects are exposed to simple "training" stimuli moving more often at high speeds than at low speeds. By subsequently testing the subjects with slow-moving stimuli of high uncertainty (low contrast), I find that their perception gradually changes in a manner consistent with the progressive acquisition of an expectation that favours progressively higher speeds. Thus subjects appear to gradually internalize the speed statistics of the stimulus ensemble over the duration of the experiment. I model these results using an existing Bayesian model of motion perception that incorporates a speed prior with a peak at zero, extending the model so that the mean gradually shifts away from zero. Although the first experiment presents evidence for the plasticity of the speed prior, the experimental paradigm and the constraints of the model limit the accuracy and precision in the reconstruction of observers’ priors. To address these limitations, I perform a different experiment where subjects compare the speed of moving gratings of different contrasts. The new paradigm allows more precise measurements of the contrast-dependent biases in perceived speed. Using a less constrained Bayesian model, I extract the priors of subjects and find considerable interindividual variability. Furthermore, noting that the Bayesian model cannot account for certain subtleties in the data, I combine the model with a non-Bayesian, physiologically motivated model of speed tuning of cortical neurons and show that the combination offers an improved description of the data. Using the paradigm of the second experiment, I then explore the role of visual experience on the form of the speed prior. By recruiting avid video gamers (who are routinely exposed to high speeds) and nongamers of both sexes, I study the differences in the prior among groups and find, surprisingly, that subjects’ speed priors depend more on gender than on gaming experience. In a final series of experiments similar to the first, I also test subjects on variations of the trained stimulus configuration – namely different orientations and motion directions. Subjects’ responses suggest that they are able to apply the changed prior to different orientations and, furthermore, that the changed prior persists for at least a week after the end of the experiment. These results provide further support for the plasticity of the speed prior but also suggest that the learned prior may be used only across similar stimulus configurations, whereas in sufficiently different configurations or contexts a “default” prior may be used instead.

Internet of Things (IoT) is becoming commonplace in people's daily life. Even, many governments' authorities have already deployed a very large number of IoT sensors toward their smart city initiative and development road-map. However, lack of semantics in the presentation of IoT-based sensory data represents the perception complexity by general people. Adding semantics to the IoT sensory data remains a challenge for smart cities and environments. In this thesis proposal, we present an implementation that provides a meaningful IoT sensory data notifications approach about indoor and outdoor environment status for people and authorities. The approach is based on analyzing spatio-temporal thresholds that compose of multiple IoT sensors readings. Our developed IoT sensory data analytics adds real-time semantics to the received sensory raw data stream by converting the IoT sensory data into meaningful and descriptive notifications about the environment status such as green locations, emergency zone, crowded places, green paths, polluted locations, etc. Our adopted IoT messaging protocol can handle a very large number of dynamically added static and dynamic IoT sensors publication and subscription processes. People can customize the notifications based on their preference or can subscribe to existing semantic notifications in order to be acknowledged of any concerned environmental condition. The thesis is supposed to come up with three contributions. The first, an IoT approach of a three-layer architecture that extracts raw sensory data measurements and converts it to a contextual-aware format that can be perceived by people. The second, an ontology that infers a semantic notification of multiple sensory data according to the appropriate spatio-temporal reasoning and description mechanism. We used a tool called Protégé to model our ontology as a common IDE to build semantic knowledge. We built our ontology through extending a well-known web ontology called Semantic Sensor Network (SSN). We built the extension from which six classes were adopted to derive our SENS-IT ontology and fulfill our objectives. The third, a fuzzy system approach is proposed to make our system much generic of providing broader semantic notifications, so it can be agile enough to accept more measurements of multiple sensory sources.

Made available in DSpace on 2014-06-11T19:32:09Z (GMT). No. of bitstreams: 0 Previous issue date: 2007-02-23Bitstream added on 2014-06-13T20:03:12Z : No. of bitstreams: 1 berti_ml_dr_assis.pdf: 530715 bytes, checksum: 3898e59d8cc563b376b6a8d3af2c6d1d (MD5) / O trabalho procura investigar os mecanismos de referenciação usados em produções de alunos de Ensino Médio para deixar pistas para que seu leitor faça inferências durante a leitura para produzir sentidos ao que lê. A reconstrução por inferenciação permite estabelecer o elo entre as informações explícitas e as implícitas no co-texto, em um determinado contexto. A inferência é uma estratégia muito importante para que se tome um texto como coeso e coerente, em termos de progressão referencial, colaborando de maneira decisiva para a produção de sentidos.A partir dos pressupostos da Lingüística textual e das teorias sobre leitura, apresenta a relação autor-texto-leitor no processo de produção e recepção do texto. Analisaram-se quantitativa e qualitativamente produções de textos de alunos de Ensino Médio nas quais se verificou o uso de seqüências pronominais, de repetições lexicais, das expressões nominais definidas, anáforas indiretas no texto ou referentes ao contexto, as quais permitem ao leitor fazer inferências e aturar como co-autor na produção dos sentidos. / The work investigates the mechanisms of reference used in High School student s productions in order to leave hints so that the reader com make inferences during the reading to produce meanings of what is being read. The reconstructions through inference permits to stablish the link among the explicit and implicit pieces of information in the co-text, in a determined context. The inference is a very important strategy to make a text cohesive and coherent, in terms of reference progression, contributing in a essencial way in the meaning production. From the textual linguistic and the theories about reading. The work presents the relationship author-text-reader in the process of text production and reception. It was analysed High School student s productions in quality and amount in which it was verified the use of pronominal sequences, of lexical repetitions, of defined nominal expressions, indirect anaphora in the text or referred to the context, wich allow the reader to make inferences and act as co-author in the production of meanings.

Perception is strongly influenced by our expectations, especially under situations of uncertainty. A growing body of work suggests that perception is akin to Bayesian Inference in which expectations can be viewed as ‘prior’ beliefs that are combined via Bayes’ rule with sensory evidence to form the ‘posterior’ beliefs. In this thesis, I aim to answer open questions regarding the nature of expectations in perception, and, in particular, what the limits of complexity and specificity in developing expectations are, and how expectations of different temporal properties develop and interact. First, I conducted a psychophysical experiment to investigate whether human observers are able to implicitly develop distinct expectations using colour as a distinguishing factor. I interleaved moving dot displays of two different colours, either red or green, with different motion direction distributions. Results showed that statistical information can transfer from one group of stimuli to another but observers are also able to learn two distinct priors under specific conditions. In a collaborative work, I implemented an online learning computational model, which showed that subjects’ behaviour was not in disagreement with a near-optimal Bayesian observer, and suggested that observers might prefer simple models which are consistent with the data over complex models. Next, I investigated experimentally whether selective manipulation of rewards can affect an observer’s perceptual performance in a similar manner to manipulating the statistical properties of stimuli. Results showed that manipulation of the reward scheme had similar effects on perception as statistical manipulations in trials where a stimulus was presented but not in the absence of stimulus. Finally, I used a novel visual search task to investigate how expectations of different timescales (from the last few trials to hours to long-term statistics of natural scenes) interact to alter perception. Results suggested that recent exposure to a stimulus resulted in significantly improved detection performance and significantly more visual ‘hallucinations’ but only at positions at which it was more probable that a stimulus would be presented. These studies provide new insights into the approximations that neural systems must make to implement Bayesian inference. Complexity does not seem to necessarily be a prohibitive factor in learning but the system also factors the provided evidence and potential gain in regards to learning complex priors and applying them in distinct contexts. Further, what aspects of the statistics of the stimuli are learned and used, and how selective attention modulates learning can crucially depend on specific task properties such as the timeframe of exposure, complexity, or the observer’s current goals and beliefs about the task.

Understanding and characterizing the processes driving social interactions is one of the fundamental problems in social network research. A particular instance of this problem, known as link prediction, has recently attracted considerable attention in various research communities. Link prediction has many important commercial applications, e.g., recommending friends in an online social network such as Facebook and suggesting interesting pins in a collection sharing network such as Pinterest. This work is focused on the temporal link prediction problem: Given a sequence of graph snapshots G1, · ··, Gt from time 1 to t, how do we predict links in future time t + 1? To perform link prediction in a network, one needs to construct models for link probabilities between pairs of nodes. A temporal latent space model is proposed that is built upon latent homophily assumption and temporal smoothness assumption. First, the proposed modeling allows to naturally incorporate the well-known homophily effect (birds of a feather flock together). Namely, each dimension of the latent space characterizes an unobservable homogeneous attribute, and shared attributes tend to create a link in a network.

No description available.

The privacy issues in online social networks (OSNs) have been increasingly arousing the public awareness since it is possible for attackers to launch several kinds of attacks to obtain users' sensitive and private information by exploiting the massive data obtained from the networks. Even if users conceal their sensitive information, attackers can infer their secrets by studying the correlations between private and public information with background knowledge. To address these issues, the thesis focuses on the inference attack and its countermeasures. First, we study how to launch the inference attack to profile OSN users via relationships and network characteristics. Due to both user privacy concerns and unformatted textual information, it is quite difficult to build a completely labeled social network directly. However, both social relations and network characteristics can help attribute inference to profile OSN users. We propose several attribute inference models based on these two factors and implement them with Naive Bayes, Decision Tree, and Logistic Regression. Also, to study network characteristics and evaluate the performance of our proposed models, we use a well-labeled Google employee social network extracted from Google+ for inferring the social roles of Google employees. The experiment results demonstrate that the proposed models are effective in social role inference with Dyadic Label Model performing the best. Second, we model the general inference attack and formulate the privacy-preserving data sharing problem to defend against the attack. The optimization problem is to maximize the users' self-disclosure utility while preserving their privacy. We propose two privacy-preserving social network data sharing methods to counter the inference attack. One is the efficient privacy-preserving disclosure algorithm (EPPD) targeting the high utility, and the other is to convert the original problem into a multi-dimensional knapsack problem (d-KP) which can be solved with a low computational complexity. We use real-world social network datasets to evaluate the performance. From the results, the proposed methods achieve a better performance when compared with the existing ones. Finally, we design a privacy protection authorization framework based on the OAuth 2.0 protocol. Many third-party services and applications have integrated the login services of popular social networking sites, such as Facebook and Google+, and acquired user information to enrich their services by requesting user's permission. However, due to the inference attack, it is still possible to infer users' secrets. Therefore, we embed our privacy-preserving data sharing algorithms in the implementation of OAuth 2.0 framework and propose RANPriv-OAuth2 to protect users' privacy from the inference attack. / Graduate

Einstein’s general theory of relativity has withstood 100 years of testing and will soon be facing one of its toughest challenges. In a few years we expect to be entering the era of the first direct observations of gravitational waves. These are tiny perturbations of space-time that are generated by accelerating matter and affect the measured distances between two points. Observations of these using the laser interferometers, which are the most sensitive length-measuring devices in the world, will allow us to test models of interactions in the strong field regime of gravity and eventually general relativity itself. I apply the tools of Bayesian inference for the examination of gravitational wave data from the LIGO and Virgo detectors. This is used for signal detection and estimation of the source parameters. I quantify the ability of a network of ground-based detectors to localise a source position on the sky for electromagnetic follow-up. Bayesian criteria are also applied to separating real signals from glitches in the detectors. These same tools and lessons can also be applied to the type of data expected from planned space-based detectors. Using simulations from the Mock LISA Data Challenges, I analyse our ability to detect and characterise both burst and continuous signals. The two seemingly different signal types will be overlapping and confused with one another for a space-based detector; my analysis shows that we will be able to separate and identify many signals present. Data sets and astrophysical models are continuously increasing in complexity. This will create an additional computational burden for performing Bayesian inference and other types of data analysis. I investigate the application of the MOPED algorithm for faster parameter estimation and data compression. I find that its shortcomings make it a less favourable candidate for further implementation. The framework of an artificial neural network is a simple model for the structure of a brain which can “learn” functional relationships between sets of inputs and outputs. I describe an algorithm developed for the training of feed-forward networks on pre-calculated data sets. The trained networks can then be used for fast prediction of outputs for new sets of inputs. After demonstrating capabilities on toy data sets, I apply the ability of the network to classifying handwritten digits from the MNIST database and measuring ellipticities of galaxies in the Mapping Dark Matter challenge. The power of neural networks for learning and rapid prediction is also useful in Bayesian inference where the likelihood function is computationally expensive. The new BAMBI algorithm is detailed, in which our network training algorithm is combined with the nested sampling algorithm MULTINEST to provide rapid Bayesian inference. Using samples from the normal inference, a network is trained on the likelihood function and eventually used in its place. This is able to provide significant increase in the speed of Bayesian inference while returning identical results. The trained networks can then be used for extremely rapid follow-up analyses with different priors, obtaining orders of magnitude of speed increase. Learning how to apply the tools of Bayesian inference for the optimal recovery of gravitational wave signals will provide the most scientific information when the first detections are made. Complementary to this, the improvement of our analysis algorithms to provide the best results in less time will make analysis of larger and more complicated models and data sets practical.