PRISM | Institutional Repository
Communities in PRISM
Select a community to browse its collections.
Recent Submissions
Annual report 2023-24, Alberta Gambling Research Institute
(Alberta Gambling Research Institute, 2024-11-22) Alberta Gambling Research Institute
Ontology-Enhanced Automated Machine Learning
(2024-11-20) Davies, Cooper T.S.; Denzinger, Jorg; Maurer, Frank; Jacob, Christian; Walker, Robert; Dick, Scott; Boyd, Jeffrey
This thesis addresses the challenge of bridging the gap between traditional Problem-Specific Machine Learning (PSML) and Automated Machine Learning (AutoML) systems. While PSML offers high accuracy but demands substantial expertise, AutoML aims to auto-mate the process of building a machine learning (ML) model but often lacks domain-specific knowledge. To address this, we propose Ontology-Enhanced AutoML, a novel approach that integrates domain knowledge from ontologies into the AutoML pipeline. We first examine the current landscape of AutoML, highlighting the complexities faced by a system in selecting appropriate algorithms and hyperparameters. We identify the limitations of existing AutoML systems, particularly their blind reliance on datasets, which often leads to poor performance and lengthy training times. Our thesis presents experiments demonstrating the effectiveness of Ontology-Enhanced AutoML in mitigating these challenges. By incorporating mechanisms for ontology-based feature extraction and example filtering, we demonstrate significant improvements in accu-racy and optimization time compared to traditional AutoML. These results highlight the potential of Ontology-Enhanced AutoML to provide a wide range of systems lying between the extremes of PSML and AutoML. This thesis contributes not only a technical solution but also a conceptual framework for understanding ML as a spectrum. We discuss implications for future research and the potential for further advancements in bridging the gap between domain expertise and ML proficiency.
Single-player to Two-player Knowledge Transfer in Atari 2600 Games
(2024-11-18) Saadat, Kimiya; Zhao, Richard; Abou-Zeid, Hatem; Aycock, John
Playing two-player games using reinforcement learning and self-play can be challenging due to the complexity of two-player environments and the potential instability in the training process. It is proposed that a reinforcement learning algorithm can train more efficiently and achieve improved performance in a two-player game by leveraging the knowledge from the single-player version of the same game. This study examines the proposed idea in ten different Atari 2600 environments using the Atari 2600 RAM as the input state. The advantages of using transfer learning from a single-player training process over training in a two-player setting from scratch are discussed, and the results are demonstrated in several metrics, such as the training time and average total reward. Finally, a method for calculating RAM complexity and its relationship to performance after transfer is discussed. Results show that in most cases transferred agent is performing better than the agent trained from scratch while taking less time to train. Moreover, it is shown that RAM complexity can be used as a weak predictor to predict the transfer's effectiveness.
Vaccine Serotypes Continue to Cause Invasive Pneumococcal Disease in the Late PCV Era Especially in Unhoused Adults
(Canadian Public Health Association, 2024-11-26) Leah J Ricketson; Shannon N Pyra; James D Kellner
Vaccine serotypes continue to cause invasive pneumococcal disease in the post-PCV era especially in unhoused adults
Introduction: PCV7 vaccine was introduced in 2002 to Calgary, and PCV13 was introduced in 2010 in Calgary. Unhoused people account for about 0.2% of the population of Calgary, but 19% of IPD cases from 2000 to 2016.
Methods: The Calgary area Streptococcus pneumoniae research (CASPER) team has been conductin population based surveillance on invasive pneumococcal disease in Calgary, Alberta since 1998. When S. pneumoniae is isolated by culture from a sterile sample, the CASPER team completes a chart review. In addition, antibiotic susceptibility testing, and serotyping through the Quellung reaction are completed by the collaborating laboratory.
Results: Following PCV7 introduction we saw a decline in serotype 4 (ST4). However, after PCV13 was introduced in 2010, ST4 began to increase in prevalence again, primarily in adults. From 2020-2022 unhoused adults accounted for over 30% of all IPD cases despite representing a small proportion of the Calgary population. ST4 represented 42% of IPD cases amongst the unhoused population in 2022, and has been a prevalent serotype in unhoused adults since 2014. The second most prevalent were ST3, ST7F, and ST19A, each accounting for 10% of cases in unhoused adults in 2022. All of these serotypes are included in PCV13 and the newly licenced PCV15 and PCV20. However, ST9V and ST4 are not included in the novel unlicenced V116 vaccine, which includes 9 serotypes that are in PCV20 and 11 unique serotypes.
Conclusions: Serotypes 4, 3, 7F and 9V are the most prevalent serotypes causing disease in Calgary in 2022/2023 in adults despite high levels of vaccination in children. The newly licenced PCV15 and PCV20 vaccines include all these serotypes, but the novel 21-valent V116 vaccine does not. Over 30% of IPD in the last three years was amongst unhoused or provisionally housed people highlighting the importance of reaching this population with vaccines.
Invasive Pneumococcal Disease (IPD) in the Late Years of the COVID-19 Pandemic in Calgary, Canada: 2022-2023
(Canadian Public Health Association, 2024-11-26) Leah J Ricketson; Shannon N Pyra; James D Kellner
Invasive pneumococcal disease in the later years of the COVID-19 pandemic in Calgary, Canada: 2022-2023 Ricketson, LJ, Kellner, JD
Background: Pneumococcal conjugate vaccines were introduced in 2002 (PCV7) and 2010 (PCV13) in Alberta. IPD declined during the COVID-19 pandemic in Calgary then increased following reduction of non-pharmaceutical interventions (NPIs). Methods: The Calgary area Streptococcus pneumoniae Epidemiology Research (CASPER) team conducts population-based surveillance on all IPD cases presenting to Calgary and area health facilities. Here we examine the changes in incidence per 3-month periods (quarters) during the early pandemic years 2020-2021 and the late-pandemic years 2022-2023 for all ages combined. Results: Late 2022 incidence increased to above pre-PCV levels. January to June 2023 had an average incidence of 4 cases per 100,000 people per quarter, which is above the average incidence of 2.5 in the same two quarters of the late-PCV13 years: 2017-2019 (IRR: 1.5, 95%CI: 1.0-2.2, P=0.0382). By Oct-Dec 2023 incidence remained above late PCV13-era levels (IRR: 1.5, 95%CI: 1.02-2.1, P=0.0317). In 2023, PCV serotypes 4 (18.9%), 9V (12.3%), 7F (9.9%), and 3 (9.9%) were most common. The rise of ST9V is notable as this was a relatively uncommon serotype in Calgary after the introduction of PCVs. Childhood vaccine rates in Calgary declined slightly, but not significantly during the COVID-19 era. Conclusions: With the reduction of NPIs, IPD increased in 2022 to above levels seen in the late-PCV era and this continued into 2023. By the end of 2023 IPD incidence remained above late vaccine-era levels. This is likely associated with increases in antecedent viral infections, particularly influenza and RSV, as childhood vaccine rates remained stable. The most common serotypes in 2023 are serotypes found in PCV13, as well as the new PCV15 and PCV20 vaccines. PCV13 serotypes persist, especially in adults and some are on the rise despite universal PCV13 use in children. This highlights the need for more effective direct protection for adults through vaccination.