Search

Conspiracy theories thrive in today’s overly saturated media landscape, where media consumers constantly are bombarded with an abundance of information from both mainstream and alternative outlets. Facing this scenario, audiences have little other choice than to somehow decide who and what to believe, and to invoke some level of effort and skill in interpreting information. This thesis seeks to explore both ‘conspiracy mentality’ (an underlying disposition toward viewing events as the product of conspiracies) and the actual endorsement of certain specific conspiracy theories in relation to several variables relevant to media consumption. Two main areas of focus are explored: a) News media trust, source preference, and literacy; and b) Critical thinking dispositions and skills. By identifying, presenting, and drawing on relevant literature, as well as evaluating former endeavors regarding measurement of specific variables, the primary aim is to quantitatively explore these relationships through answering eight research questions. A quantitative, fixed-response cross-sectional questionnaire containing primarily a combination of previously validated-, existing but altered-, as well as self-developed multi-item Likert measuring scales forms the basis of findings. Data is collected from a demographically diverse non-probability convenience sample selection of Norwegians (N=125) of all ages, socio-cultural backgrounds, and with varying political views – as well as media preferences and habits. Initial observations reveal a moderately high level of general conspiracy mentality within the sample – as well as a prevailing presence of respondents who endorse various conspiracy theories. A series of bivariate correlation analyses and one-way analyses of variance (ANOVA’s) are run based on the gathered data in order to answer all research questions. Findings suggest that high skepticism towards mainstream media, as well as a positive attitude towards – and frequent consumption of – ‘alternative’ news media are all significantly related to both conspiracy mentality and conspiracy theory endorsement. Findings further suggest that susceptibility to informal fallacious arguments, in addition to an inability to identify nonsensical, but seemingly meaningful ideas (‘pseudo-profound bullshit’), are also correlated with conspiracy theory endorsement and -mentality. Partial, albeit weak, relationships to some aspects of news media literacy – such as understanding how filtering of information in the media can affect perceptions of reality – are also discovered in relation to conspiracy theory endorsement. Surprisingly, no significant findings are uncovered in relation to critical thinking disposition or source criticism abilities; although it is hypothesized that methodological flaws may be at fault for some of these results. Some possible implications of findings for journalists and media workers are discussed where applicable throughout, and upon conclusion.

This paper describes two experiments that first explore the potential role of Chinese character writing in character visual recognition, and then examine different evaluative responses towards the quality of pinyin and character handwriting. Taken together, the results suggest that drawing Chinese characters privileges them in memory in a way that facilitates their subsequent visual recognition. This is true even when the congruency of the recognition response and other potential confounds are controlled for. In terms of the writing quality, the reader’s empathy effect can be found for handwritten characters but not pinyin, since the handwritten characters tended to be rated more highly than pinyin. The experience of an evaluator also has an impact on the evaluation of writing quality. The pedagogical implications for Chinese as a foreign language (CFL) are highlighted at the end of the paper, in particular those relating to curriculum design and teacher training.

Fractal patterns offer one way to represent the rough complexity of the natural world. Whilst they dominate many of our visual experiences in nature, little large-scale perceptual research has been done to explore how we respond aesthetically to these patterns. Previous research (Taylor et al., 2011) suggests that the fractal patterns with mid-range fractal dimensions have universal aesthetic appeal. Perceptual and aesthetic responses to visual complexity have been more varied with findings suggesting both linear (Forsythe et al., 2011) and curvilinear (Berlyne, 1970) relationships. Individual differences have been found to account for many of the differences we see in aesthetic responses but some, such as culture, have received little attention within the fractal and complexity research fields. This 2-study paper aims to test preference responses to fractal dimension and visual complexity, using a large cohort (N=443) of participants from around the world to allow universality claims to be tested. It explores the extent to which age, culture and gender can predict our preferences for fractally complex patterns. Following exploratory analysis that found strong correlations between fractal dimension and visual complexity, a series of linear mixed-effect models were implemented to explore if each of the individual variables could predict preference. The first tested a linear complexity model (likelihood of selecting the more complex image from the pair of images) and the second a mid-range fractal dimension model (likelihood of selecting an image within mid-range). Results show that individual differences can reliably predict preferences for complexity across culture, gender and age. However, in fitting with current findings the mid-range models show greater consistency in preference not mediated by gender, age or culture. This paper supports the established theory that the mid-range fractal patterns appear to be a universal construct underlying preference but also highlights the fragility of universal claims by demonstrating individual differences in preference for the interrelated concept of visual complexity. This highlights a current stalemate in the field of empirical aesthetics.

This paper describes two experiments that explore the potential role of Chinese character writing on their visual recognition. Taken together, the results suggest that drawing Chinese characters privileges them in memory in a way that facilitates their subsequent visual recognition. This is true even when the congruency of the recognition response and other potential confounds are controlled for.

Until recently, filming has been an analogue process; it requires a mechanical process to record and view, and the source material itself is prone to decay & abrasion [1]. Film is expensive to store, and prohibitively expensive to restore. All footage - historical, documentary or entertainment – may completely degrade over time. While many archival films stocks are currently being scanned and further damage thus prevented, the digital copies are far from the quality of the original. The types of aberrations found are varied, from frame jitter and line scratches to dirt and sparkle. It is the detection of the latter two (which are frame based abnormalities) that will be examined here.

This paper documents a study, carried out in the academic year 2003-2004, on fifteen factors that may influence performance on a first year object-oriented programming module. The factors included prior academic experience, prior computer experience, self-perception of programming performance and comfort level on the module and specific cognitive skills. The study found that a student's perception of their understanding of the module had the strongest correlation with programming performance, r=0.76, p‹0.01. In addition, Leaving Certificate (LC) mathematics and science scores were shown to have a strong correlation with performance. A regression module, based upon a student's perception of their understanding of the module, gender, LC mathematics score and comfort level was able to account for 79% of the variance in programming performance results.

This paper documents a study, carried out in the academic year 2004-2005, on the role of motivation and comfort-level in a first year object-oriented programming module. The study found that intrinsic motivation had a strong correlation with programming performance as did self-efficacy for learning and performance, r=0.512, p < 0.01 and r=0.567, p < 0.01 respectively. Aspects of comfort level were found to have significant correlations with performance with an instrument on programming-esteem rendering the most interesting results. A regression model based upon these factors was able to account for 60% of the variance in programming performance results.

This paper proposes the use of neural networks as a tool for studying navigation within virtual worlds. Results indicate that the network learned to predict the next step for a given trajectory. The analysis of hidden layer shows that the network was able to differentiate between two groups of users identified on the basis of their performance for a spatial task. Time series analysis of hidden node activation values and input vectors suggested that certain hidden units become specialised for place and heading, respectively. The benefits of this approach and the possibility of extending the methodology to the study of navigation in Human Computer Interaction applications are discussed. 8.7.2013 SB.

The goal of this paper is to propose a candidate for consideration as a computational principle for cognitive development in autonomous robots. The candidate in question is the theory of Cortical Software Re-Use (CSRU) and we will make the case in this paper that it provides a mechanism for the incremental construction of cognitive and language systems from simpler sensory-motor components.

We present a general event-driven algorithm for the efficient simulation of spiking neural networks. We focus in this paper on its application to self-organizing maps. Standard event-driven approaches to simulation can significantly reduce computational time, but only when network activity is relatively low. In this article, we propose several strategies to manage efficiently, large numbers of spiking events. The simulation scales well with the increase of the neural activity and is more biologically plausible than competing methods. Key-Words: - Event-driven simulation, spiking neurons, self-organizing maps.

We introduce a novel two-stage automatic XML mark-up system, which combines the WEBSOM approach to document categorisation in conjunction with the C5 inductive learning algorithm. The WEBSOM method clusters the XML marked-up documents such that semantically similar documents lie close together on a Self-Organising Map (SOM). The C5 algorithm automatically learns and applies mark-up rules derived from the nearest SOM neighbours of an unmarked document. The system learns from mark-up errors to improve accuracy. The automatically marked-up documents produced by the system are also categorized on the Self-Organizing Map, to further refine SOM's document coverage.

A connectionist attentional-shift model of eye movement control (CASMEC) in reading is described. The model provides an integrated account of a range of saccadic control effects found in reading, such as word-skipping, refixation, and of course normal saccadic progression.

A technique is described that permits the on-line construction and dynamic modification of parse trees during the processing of sentence-like-input. The approach is a combination of simple recurrent network (SRN) and recursive auto-associative memory (RAAM). The parsing technique involves teaching the SRN to build RAAM representations as it processes its input item by item. The approach is a potential component of a larger connectionist natural language processing system, and could also be used as a tool in the cognitive modelling of language understanding. Unfortunately, the modified SRN demonstrates a limited capacity for generalization.

We present an analysis of Ecommerce clickstream data using Recurrent\ud Neural Networks (RNN), Gated Recurrent Units (GRU) and\ud Long-Short Term Memory (LSTM). Our analysis highlights the\ud substantial difference in the predictive power of LSTM models depending\ud on whether or not hidden state is shared across batches and\ud also assesses the ability of RNNs to learn and use both session-local\ud and dataset-global information under different sampling strategies.\ud We propose random sampling combined with stateless LSTM for\ud optimal performance of LSTM in an Ecommerce domain.

We present a neural network for predicting purchasing intent in an Ecommerce setting. Our main contribution is to address the significant investment in feature engineering that is usually associated with state-of-the-art methods such as Gradient Boosted Machines. We use trainable vector spaces to model varied, semi-structured input data comprising categoricals, quantities and unique instances. Multi-layer recurrent neural networks capture both session-local and dataset-global event dependencies and relationships for user sessions of any length. An exploration of model design decisions including parameter sharing and skip connections further increase model accuracy. Results on benchmark datasets deliver classification accuracy within 98% of state-of-the-art on one and exceed state-of-the-art on the second without the need for any domain / dataset-specific feature engineering on both short and long event sequences., Comment: Accepted to SIGIR eCom workshop, Ann Arbor, Michigan, USA, 2018