Effective mental workload measurement is critical because mental workload significantly affects human performance. A non-invasive and objective workload measurement tool is needed to overcome limitations of current mental workload measures. Further, training/learning increases mental workload during skill or knowledge acquisition, followed by a decreased mental workload, though sufficient training times are unknown. The objectives of this study were to: (1) investigate the efficacy of using thermography as a non-contact physiological measure to quantify mental workload, (2) quantify and describe the relationship between mental workload and learning/training, and, (3) introduce a method to determine a sufficient training time and an optimal human performance level for a novel task by using thermography. Three studies were conducted to address these objectives. The first study investigated the efficacy of using thermography to quantity the relationship between mental workload and facial temperature changes while learning an alpha-numeric task. Thermography measured and quantified the mental workload level successfully. Strong and significant correlations were found among thermography, performance, and subjective workload measures (MCH and SWAT ratings). The second study investigated the utility of using a psychophysical approach to determine workload levels that maximize performance on a cognitive task. The second study consisted of an adjustment session (participants adjusted their own workload levels) and work session (participants worked at the chosen workload level). Participants were found to fall into two performance groups (low and high performers by accuracy rate) and results were significantly different. Thermography demonstrated whether both group found their optimal workload level. The last study investigated efficacy of using thermography to quantify mental workload level in a complex training/learning environment. Experienced drivers’ performance data was used as criteria to indicate whether novice drivers mastered the driving skills. Strong and significant correlations were found among thermography, subjective workload measures, and performance measures in novice drivers. This study verified that thermography is a reliable and valid way to measure workload as a non-invasive and objective method. Also, thermography provided more practical results than subjective workload measures for simple and complex cognitive tasks. Thermography showed the capability to identify a sufficient training time for simple or complex cognitive tasks.

The distributed nature of Federated Learning (FL) creates security-related vulnerabilities including training-time attacks. Recently, it has been shown that well-known Byzantine-resilient aggregation schemes are indeed vulnerable to an informed adversary who has access to the aggregation scheme and updates sent by clients. Therefore, it is a significant challenge to establish successful defense mechanisms against such an adversary. To the best of our knowledge, most current aggregators are immune to single or partial attacks and none of them is expandable to defend against new attacks. We frame the robust distributed learning problem as a game between a server and an adversary that tailors training-time attacks. We introduce RobustTailor, a simulation-based algorithm that prevents the adversary from being omniscient. RobustTailor is a mixed strategy and has good expandability for any deterministic Byzantine-resilient algorithm. Under a challenging setting with information asymmetry between two players, we show that our method enjoys theoretical guarantees in terms of regret bounds. RobustTailor preserves almost the same privacy guarantees as standard FL and robust aggregation schemes. Simulation improves robustness to training-time attacks significantly. Empirical results under challenging attacks validate our theory and show that RobustTailor preforms similar to an upper bound which assumes the server has perfect knowledge of all honest clients over the course of training. / Den distribuerade karaktären hos federerade maskininlärnings-system gör dem sårbara för cyberattacker, speciellt under tiden då systemen tränas. Nyligen har det visats att många existerande Byzantine-resilienta aggregeringssystem är sårbara för attacker från en välinformerad motståndare som har tillgång till aggregeringssystemet och uppdateringarna som skickas av klienterna. Det är därför en stor utmaning att skapa framgångsrika försvarsmekanismer mot en sådan motståndare. Såvitt vi vet är de flesta nuvarande aggregatorer immuna mot enstaka eller partiella attacker och ingen av dem kan på ett enkelt sätt utvidgas för att försvara sig mot nya attacker. Vi utformar det robusta distribuerade inlärningsproblemet som ett spel mellan en server och en motståndare som skräddarsyr attacker under träningstiden. Vi introducerar RobustTailor, en simuleringsbaserad algoritm som förhindrar att motståndaren är allvetande. RobustTailor är en blandad strategi med god expanderbarhet för alla deterministiska Byzantine-resilienta algoritmer. I en utmanande miljö med informationsasymmetri mellan de två spelarna visar vi att vår metod har teoretiska garantier i form av gränser för ånger. RobustTailor har nästan samma integritetsgarantier som standardiserade federerade inlärnings- och robusta aggregeringssystem. Vi illustrerar även hur simulering förbättrar robustheten mot attacker under träningstiden avsevärt. Empiriska resultat vid utmanande attacker bekräftar vår teori och visar att RobustTailor presterar på samma sätt som en övre gräns som förutsätter att servern har perfekt kunskap om alla ärliga klienter under utbildningens gång.

Deep Neural Networks (DNNs) are becoming increasingly important in various aspects of human life, particularly in safety-critical areas such as autonomous driving and aerospace systems. However, soft errors including bit-flips can significantly impact the performance of these systems, leading to serious consequences. To ensure the reliability of DNNs, it is essential to guarantee their performances. Many solutions have been proposed to enhance the trustworthiness of DNNs, including traditional methods like error correcting code (ECC) that can mitigate and detect soft errors but come at a high cost of redundancy. This thesis proposes a new method of correcting soft errors in DNNs using Deep Reinforcement Learning (DRL) and Transfer Learning (TL). DRL agent can learn the knowledge of identifying the layer-wise critical weights of a DNN. To accelerate the training time, TL is used to apply this knowledge to train other layers. The primary objective of this method is to ensure acceptable performance of a DNN by mitigating the impact of errors on it while maintaining low redundancy. As a case study, we tested the proposed method approach on a multilayer perception (MLP) and ResNet-18, and our results show that our method can save around 25% redundancy compared to the baseline method ECC while achieving the same level of performance. With the same redundancy, our approach can boost system performance by up to twice that of conventional methods. By implementing TL, the training time of MLP is shortened to around 81.11%, and that of ResNet-18 is shortened to around 57.75%. / DNNs blir allt viktigare i olika aspekter av mänskligt liv, särskilt inom säkerhetskritiska områden som autonom körning och flygsystem. Mjuka fel inklusive bit-flip kan dock påverka prestandan hos dessa system avsevärt, vilket leder till allvarliga konsekvenser. För att säkerställa tillförlitligheten hos DNNs är det viktigt att garantera deras prestanda. Många lösningar har föreslagits för att förbättra tillförlitligheten för DNNs, inklusive traditionella metoder som ECC som kan mildra och upptäcka mjuka fel men som har en hög kostnad för redundans. Denna avhandling föreslår en ny metod för att korrigera mjuka fel i DNN med DRL och TL. DRL-agenten kan lära sig kunskapen om att identifiera de lagermässiga kritiska vikterna för en DNN. För att påskynda träningstiden används TL för att tillämpa denna kunskap för att träna andra lager. Det primära syftet med denna metod är att säkerställa acceptabel prestanda för en DNN genom att mildra inverkan av fel på den samtidigt som låg redundans bibehålls. Som en fallstudie testade vi den föreslagna metodmetoden på en MLP och ResNet-18, och våra resultat visar att vår metod kan spara cirka 25% redundans jämfört med baslinjemetoden ECC samtidigt som vi uppnår samma prestationsnivå. Med samma redundans kan vårt tillvägagångssätt öka systemets prestanda med upp till dubbelt så högt som för konventionella metoder. Genom att implementera TL förkortas träningstiden för MLP till cirka 81.11%, och den för ResNet-18 förkortas till cirka 57.75%.

<p>This research was a pilot study in a larger project that focused on how to retrieve knowledge from retiring long-term employees of a small university utility plant, incorporate that material into their existing training program, and during the process reduce the training time for current and future employees. Wade utility plant faced the retirement of eight employees with nearly 200 years of corporate knowledge within three years, but their current training program required seven to nine years to complete. The study utilized interviews, first-hand observation and partnership with current employees to explore how best to obtain the corporate knowledge that would be lost when the proletarian workers retired. The study revealed that the training program needed to be updated, and communication, trust and training evaluation continuity needed to be addressed. Due to these issues, trust was built through transparency by the researcher, and suggestions were made to management for moving forward. This study adds to the body of knowledge by utilizing knowledge capture techniques in a utility plant, highlighting effective knowledge capture techniques for proletarian workers, the importance of corporate planning for the effect of group retirements, and how incorporating proletarian workers into training creation can make a positive impact on company relationships.</p>

Organizations use Business Intelligence (BI) systems to monitor how well they are meeting their goals and objectives. Yet, very often BI systems do not include clear models of the organization’s goals or of how to measure whether they are satisfied or not. Several researchers now attempt to integrate goal models into BI systems, but there are still major challenges related to how to get access to the BI data to populate the part of the goal model (often indicators) used to assess goal satisfaction. This thesis explores a new approach to integrate BI systems with goal models. In particular, it explores the integration of IBM Cognos Business Intelligence, a leading BI tool, with an Eclipse-based goal modeling tool named jUCMNav. jUCMNav is an open source graphical editor for the User Requirements Notation (URN), which includes the Use Case Map notation for scenarios and processes and the Goal-oriented Requirement Language for business objectives. URN was recently extended with the concept of Key Performance Indicator (KPI) to enable performance assessment and monitoring of business processes. In jUCMNav, KPIs are currently calculated or modified manually. The new integration proposed in this thesis maps these KPIs to report elements that are generated automatically by Cognos based on the model defined in jUCMNav at runtime, with minimum effort. We are using IBM Cognos Mashup Service, which includes web services that enable the retrieval of report elements at the most granular level. This transformation provides managers and analysts with useful goal-oriented and process-oriented monitoring views fed by just-in-time BI information. This new solution also automates retrieving data from Cognos servers, which helps reducing the high costs usually caused by the amount of manual work required otherwise. The novel approach presented in this thesis avoids manual report generation and minimizes any contract with respect to the location of manually created reports, hence leading to better usability and performance. The approach and its tool support are illustrated with an ongoing example, validated with a case study, and verified through testing.

Organizations use Business Intelligence (BI) systems to monitor how well they are meeting their goals and objectives. Yet, very often BI systems do not include clear models of the organization’s goals or of how to measure whether they are satisfied or not. Several researchers now attempt to integrate goal models into BI systems, but there are still major challenges related to how to get access to the BI data to populate the part of the goal model (often indicators) used to assess goal satisfaction. This thesis explores a new approach to integrate BI systems with goal models. In particular, it explores the integration of IBM Cognos Business Intelligence, a leading BI tool, with an Eclipse-based goal modeling tool named jUCMNav. jUCMNav is an open source graphical editor for the User Requirements Notation (URN), which includes the Use Case Map notation for scenarios and processes and the Goal-oriented Requirement Language for business objectives. URN was recently extended with the concept of Key Performance Indicator (KPI) to enable performance assessment and monitoring of business processes. In jUCMNav, KPIs are currently calculated or modified manually. The new integration proposed in this thesis maps these KPIs to report elements that are generated automatically by Cognos based on the model defined in jUCMNav at runtime, with minimum effort. We are using IBM Cognos Mashup Service, which includes web services that enable the retrieval of report elements at the most granular level. This transformation provides managers and analysts with useful goal-oriented and process-oriented monitoring views fed by just-in-time BI information. This new solution also automates retrieving data from Cognos servers, which helps reducing the high costs usually caused by the amount of manual work required otherwise. The novel approach presented in this thesis avoids manual report generation and minimizes any contract with respect to the location of manually created reports, hence leading to better usability and performance. The approach and its tool support are illustrated with an ongoing example, validated with a case study, and verified through testing.

Organizations use Business Intelligence (BI) systems to monitor how well they are meeting their goals and objectives. Yet, very often BI systems do not include clear models of the organization’s goals or of how to measure whether they are satisfied or not. Several researchers now attempt to integrate goal models into BI systems, but there are still major challenges related to how to get access to the BI data to populate the part of the goal model (often indicators) used to assess goal satisfaction. This thesis explores a new approach to integrate BI systems with goal models. In particular, it explores the integration of IBM Cognos Business Intelligence, a leading BI tool, with an Eclipse-based goal modeling tool named jUCMNav. jUCMNav is an open source graphical editor for the User Requirements Notation (URN), which includes the Use Case Map notation for scenarios and processes and the Goal-oriented Requirement Language for business objectives. URN was recently extended with the concept of Key Performance Indicator (KPI) to enable performance assessment and monitoring of business processes. In jUCMNav, KPIs are currently calculated or modified manually. The new integration proposed in this thesis maps these KPIs to report elements that are generated automatically by Cognos based on the model defined in jUCMNav at runtime, with minimum effort. We are using IBM Cognos Mashup Service, which includes web services that enable the retrieval of report elements at the most granular level. This transformation provides managers and analysts with useful goal-oriented and process-oriented monitoring views fed by just-in-time BI information. This new solution also automates retrieving data from Cognos servers, which helps reducing the high costs usually caused by the amount of manual work required otherwise. The novel approach presented in this thesis avoids manual report generation and minimizes any contract with respect to the location of manually created reports, hence leading to better usability and performance. The approach and its tool support are illustrated with an ongoing example, validated with a case study, and verified through testing.