Electronic Theses and Dissertations (Masters)

Permanent URI for this collection

https://hdl.handle.net/10539/37969

Browse

Browsing Electronic Theses and Dissertations (Masters) by Issue Date

Filter results by year or month
Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Model Propagation for High-Parallelism in Data Compression
    (University of the Witwatersrand, Johannesburg, 2023-10) Lin, Shaw Chian; Cheng, Ling
    Recent data compression research focuses on the parallelisation of existing algorithms (LZ77, BZIP2 etc.) by exploiting their inherent parallelism. Little work has been performed on parallelising highly sequential algorithms, whose slow compression speeds would benefit the most from parallelism. This dissertation presents a generalised parallelisation approach that can be potentially adopted by any compression algorithms, with model sequentiality in mind. The scheme presents a novel divide-and-conquer approach when dividing the data stream into smaller data blocks for parallelisation. The scheme, branching propagation, is implemented with prediction by partial matching (PPM), an algorithm of the statistical-modelling family known for their serial nature, which is shown to suffer from compression ratio increases when parallelised. A speedup of 5.2-7x has been achieved at 16 threads, with at most a 6.5% increase in size relative to serial performance, while the conventional approach showed up to a 7.5x speedup with an 8.0% increase. The branching propagation approach has been shown to offer better compression ratios over conventional approaches with increasing parallelism (a difference of 11% increase at 256 threads), albeit at slightly slower speeds. To quantify the speedup over ratio penalty, an alternate metric called speedup-to-ratio increase (SRI) is used. This shows that when serial dependency is maintained, branching propagation is superior in standard configurations, which offers substantial speed while minimising the compression ratio penalty relative to the speedup. However, at lower serial dependency, the conventional approach is generally preferable, with 9-16x speedup per 1% increase in compression ratio at maximal speed compared to branching propagation’s 6-13x speedup per 1%.
  • Thumbnail Image
    Item
    Evaluation and algorithmic adaptation of brain state control through audio entertainment
    (University of the Witwatersrand, Johannesburg, 2023-12) Cassim, Muhammed Rashaad; Rubin, David; Pantanowitz, Adam
    This dissertation presents the design and evaluation of a system that can alter the dominant brain state of participants through audio entrainment. The ‘rch broadly aimed to identify the possible improvements of a dynamic entrainment stimulus when compared to a set entrainment stimulus. The dynamic entrainment stimulus was controlled by a Q-Learning (QL) model. The experiment sought to build on previous research by implementing existing entrainment methods in Virtual Reality and dynamically optimising the entrainment stimulus. The neurological effects of the stimuli were evaluated by analysing electroencephalogram measurements. It was found that a set 24 Hz entrainment stimulus increased the power of Beta band brain waves relative to a control condition. Further, contrary to existing research, it was found that the entrainment stimulus did not have a notable effect on brainwave connectivity at the entrainment frequency. The study subsequently evaluated if the QL agent could learn to optimise the entrainment stimulus. The agent was allowed to switch between an 18 and 24 Hz entrainment stimulus and succeeded in learning an optimised policy. The QL driven stimulus yielded results that generally exhibited the same characteristics as the set entrainment stimulus when using power and connectivity analysis methods. Furthermore, the power analysis indicated that the QL driven stimulus was able to affect a broader range of frequencies within the targeted band. The QL driven stimulus, additionally, resulted in higher meta-analysis metric values in some aspects. These factors indicate that it was able to have a more consistent impact on targeted brain waves. Lastly, results from participants whose stimulus was controlled by a QL driven stimulus using optimal actions indicated that the optimised actions created a more sustained increase in Beta band activity when compared to any other results, indicating the impact of the optimised policy learned.