<p dir="ltr">This research explored a solution to a high accident rate in remotely operating Unmanned Aerial Vehicles (UAVs) in a complex environment; it presented a new Brain-Computer Interface (BCI) enabled supervisory control system to fuse human and machine intelligence seamlessly. This study was highly motivated by the critical need to enhance the safety and reliability of UAV operations, where accidents often stemmed from human errors during manual controls. Existing BCIs confronted the challenge of trading off a fully remote control by humans and an automated control by computers. This study met such a challenge with the proposed supervisory control system to optimize human-machine collaboration, prioritizing safety, adaptability, and precision in operation.</p><p dir="ltr">The research work included designing, training, and testing BCI and the BCI-enabled control system. It was customized to control a UAV where the user’s motion intents and cognitive states were monitored to implement hybrid human and machine controls. The DJI Tello drone was used as an intelligent machine to illustrate the application of the proposed control system and evaluate its effectiveness through two case studies. The first case study was designed to train a subject and assess the confidence level for BCI in capturing and classifying the subject’s motion intents. The second case study illustrated the application of BCI in controlling the drone to fulfill its missions.</p><p dir="ltr">The proposed supervisory control system was at the forefront of cognitive state monitoring to leverage the power of an ML model. This model was innovative compared to conventional methods in that it could capture complicated patterns within raw EEG data and make decisions to adopt an ensemble learning strategy with the XGBoost. One of the key innovations was capturing the user’s intents and interpreting these into control commands using the EmotivBCI app. Despite the headset's predefined set of detectable features, the system could train the user’s mind to generate control commands for all six degrees of freedom of adapting to the quadcopter by creatively combining and extending mental commands, particularly in the context of the Yaw rotation. This strategic manipulation of commands showcased the system's flexibility in accommodating the intricate control requirements of an automated machine.</p><p dir="ltr">Another innovation of the proposed system was its real-time adaptability. The supervisory control system continuously monitors the user's cognitive state, allowing instantaneous adjustments in response to changing conditions. This innovation ensured that the control system was responsive to the user’s intent and adept at prioritizing safety through the arbitrating mechanism when necessary.</p>

Due to the large share of production costs, the lifespan of an electric vehicle’s (EV) lithium-ion traction battery should be as long as possible. The optimisation of the EV’s operating strategy with regard to battery life requires a regular evaluation of the battery’s state-of-health (SOH). Yet the SOH, the remaining battery capacity, cannot be measured directly through sensors but requires the elaborate conduction of special characterisation tests. Considering the limited number of test facilities as well as the rapidly growing number of EVs, time-efficient and scalable SOH estimation methods are urgently needed and are the object of investigation in this work. The developed virtual SOH experiment originates from the incremental capacity measurement and solely relies on the commonly logged battery management system (BMS) signals to train the digital battery twins. The first examined dataset with identical load profiles for new and aged battery state serves as proof of concept. The successful SOH estimation based on the second dataset that consists of varying load profiles with increased complexity constitutes a step towards the application on real driving cycles. Assuming that the load cycles contain pauses and start from the fully charged battery state, the SOH estimation succeeds either through a steady shift of the load sequences (variant one) with an average deviation of 0.36% or by random alignment of the dataset’s subsequences (variant two) with 1.04%. In contrast to continuous capacity tests, the presented framework does not impose restrictions to small currents. It is entirely independent of the prevailing and unknown ageing condition due to the application of battery models based on the novel encoder–decoder architecture and thus provides the cornerstone for a scalable and robust estimation of battery capacity on a pure data basis.

Sensitive data leaking from a system can have tremendous negative consequences, such as discrimination, social stigma, and fraudulent economic consequences for those whose data has been leaked. Therefore, it’s of utmost importance that sensitive data is not leaked from a system. This thesis investigated different methods to prevent sensitive patient data from leaking in a machine learning system. Various methods have been investigated and evaluated based on previous research; the methods used in this thesis are a large language model (LLM) for code analysis and a membership inference attack on models to test their privacy level. The LLM code analysis results show that the Llama 3 (an LLM) model had an accuracy of 90% in identifying malicious code that attempts to steal sensitive patient data. The model analysis can evaluate and determine membership inference of sensitive patient data used for training in machine learning models, which is essential for determining data leakage a machine learning model can pose in machine learning systems. Further studies in increasing the deterministic and formatting of the LLM‘s responses must be investigated to ensure the robustness of the security system that utilizes LLMs before it can be deployed in a production environment. Further studies of the model analysis can apply a wider variety of evaluations, such as increased size of machine learning model types and increased range of attack testing types of machine learning models, which can be implemented into machine learning systems. / Känsliga data som läcker från ett system kan ha enorma negativa konsekvenser, såsom diskriminering, social stigmatisering och negativa ekonomiska konsekvenser för dem vars data har läckt ut. Därför är det av yttersta vikt att känsliga data inte läcker från ett system. Denna avhandling undersökte olika metoder för att förhindra att känsliga patientdata läcker ut ur ett maskininlärningssystem. Olika metoder har undersökts och utvärderats baserat på tidigare forskning; metoderna som användes i denna avhandling är en stor språkmodell (LLM) för kodanalys och en medlemskapsinfiltrationsattack på maskininlärnings (ML) modeller för att testa modellernas integritetsnivå. Kodanalysresultaten från LLM visar att modellen Llama 3 hade en noggrannhet på 90% i att identifiera skadlig kod som försöker stjäla känsliga patientdata. Modellanalysen kan utvärdera och bestämma medlemskap av känsliga patientdata som används för träning i maskininlärningsmodeller, vilket är avgörande för att bestämma den dataläckage som en maskininlärningsmodell kan exponera. Ytterligare studier för att öka determinismen och formateringen av LLM:s svar måste undersökas för att säkerställa robustheten i säkerhetssystemet som använder LLM:er innan det kan driftsättas i en produktionsmiljö. Vidare studier av modellanalysen kan tillämpa ytterligare bredd av utvärderingar, såsom ökad storlek på maskininlärningsmodelltyper och ökat utbud av attacktesttyper av maskininlärningsmodeller som kan implementeras i maskininlärningssystem.

Trots de framsteg som gjorts inom artificiell intelligens (AI) och maskininlärning (ML), uppkommer utmaningar gällande deras miljöpåverkan. Fokuset på att skapa avancerade och träffsäkra modeller innebär ofta att omfattande beräkningsresurser krävs, vilket leder till en hög energiförbrukning. Syftet med detta arbete är att undersöka ämnet grön AI och sambandet mellan prestanda och energiförbrukning hos två ML-algoritmer. De algoritmer som undersöks är beslutsträd och stödvektormaskin (SVM), med hjälp av två dataset: Bank Marketing och MNIST. Prestandan mäts med utvärderingsmåtten noggrannhet, precision, recall och F1-poäng, medan energiförbrukningen mäts med verktyget Intel VTune Profiler. Arbetets resultat visar att en högre prestanda resulterade i en högre energiförbrukning, där SVM presterade bäst men också förbrukade mest energi i samtliga tester. Vidare visar resultatet att optimering av modellerna resulterade både i en förbättrad prestanda men också i en ökad energiförbrukning. Samma resultat kunde ses när ett större dataset användes. Arbetet anses inte bidra med resultat eller riktlinjer som går att generalisera till andra arbeten. Däremot bidrar arbetet med en förståelse och medvetenhet kring miljöaspekterna gällande AI, vilket kan användas som en grund för att undersöka ämnet vidare. Genom en ökad medvetenhet kan ett gemensamt ansvar tas för att utveckla AI-lösningar som inte bara är kraftfulla och effektiva, utan också hållbara. / Despite the advancements made in artificial intelligence (AI) and machine learning (ML), challenges regarding their environmental impact arise. The focus on creating advanced and accurate models often requires extensive computational resources, leading to a high energy consumption. The purpose of this work is to explore the topic of green AI and the relationship between performance and energy consumption of two ML algorithms. The algorithms being evaluated are decision trees and support vector machines (SVM), using two datasets: Bank Marketing and MNIST. Performance is measured using the evaluation metrics accuracy, precision, recall, and F1-score, while energy consumption is measured using the Intel VTune Profiler tool. The results show that higher performance resulted in higher energy consumption, with SVM performing the best but also consuming the most energy in all tests. Furthermore, the results show that optimizing the models resulted in both improved performance and increased energy consumption. The same results were observed when a larger dataset was used. This work is not considered to provide results or guidelines that can be generalized to other studies. However, it contributes to an understanding and awareness of the environmental aspects of AI, which can serve as a foundation for further exploration of the topic. Through increased awareness, shared responsibility can be taken to develop AI solutions that are not only powerful and efficient but also sustainable.

Teknik och fastighetsförvaltningen är en förvaltning inom Västerås stad som ansvarar för byggandet av Västerås stad. Förvaltningen är intresserad av att få en bredare kunskap om optimering av underhållssystem för luftkvalitet och hur det skulle leda till energibesparing. Uppkomsten till deras intresse för om optimering av underhållssystem för luftkvalitet och energibesparing, är av anledning att de söker nya innovativa möjligheter att optimera luftkvalitet inom deras befintliga och nya fastigheter inom Västerås stads kommun. Projektgruppen samt teknik och fastighetsförvaltningen valde att lägga fokus på Hamreskolan där de i dagsläget har ett gediget underhållssystem för luftkvaliteten men har en önskan till förbättring. Skälet är deras upplevelse av luftkvalitet som inte är optimal, upplevelsen är att man känner sig trött, att det är kallt och kvavt ibland även för varmt inne i lokalerna. Bra luftkvalite är väsentligt för det påverkar både personalen och eleverna prestationsförmåga prioriterades detta. Målet med detta examensarbete är att presentera förbättringsförslag för att optimera underhållssystemet i Hamreskolan. Underhållssystemet innefattar ventilationssystemet och styrsystemet där dess syfte är att underhålla luftkvaliteten. De metoder som användes för framtagandet av förbättrings förslagen är djup litteraturstudie, platsbesök i Hamreskolan, brainstorming med förvaltare från Teknik och fastighetsförvaltningen samt pugh matris för validering av förbättrings förslagen. I detta examensarbete presenteras och diskuteras de förbättringsförslag som kommer medföra positiva effekter för Hamreskolan vid implementation. Dessa förbättringsförslag behövs inte nödvändigtvist begränsas till endast implementation vid Hamreskolan, det går även att implementera vid flera fastigheter inom Västerås stad, Teknik och fastighetsförvaltning. Vid utvecklande av förbättringsförslagen har realitet för funktionalitet och dess effekt vid implementation i Hamreskolan varit i åtanken. / Technology and property management is an administration within the city of Västerås that is responsible for the construction of the city of Västerås. The administration is interested in gaining a broader knowledge of optimizing maintenance systems for air quality and how it would lead to energy savings. The emergence of their interest in optimizing maintenance systems for air quality and energy savings, is due to seeking new innovative opportunities to optimize air quality within their existing and new properties within the City of Västerås. The project group as well as technology and property management chose to focus on Hamreskolan, where they currently have a solid maintenance system for air quality but have a desire for improvement. The reason is their experience of air quality which is not optimal, the experience is that you feel tired, that it is cold and sometimes even too hot inside the premises. Good air quality is essential because it affects both the staff and the student's performance priorities. The aim of this thesis is to present improvement proposals to optimize the maintenance system in Hamreskolan. The maintenance system includes the ventilation system and the control system where its purpose is to maintain the air quality. The methods used for the preparation of improvement proposals are in-depth literature study, site visits to Hamreskolan, brainstorming with managers from Technology and Property Management and a pugh matrix for validation of improvement proposals. In this thesis, the improvement proposals that will have positive effects for Hamreskolan upon implementation are presented and discussed. These improvement proposals do not necessarily have to be limited to only implementation at Hamreskolan, it is also possible to implement at several properties within the City of Västerås, Technology and property management. In developing the improvement proposals, the reality for functionality and its effect when implemented in Hamreskolan has been in mind.

Denna studie undersöker användningen av maskininlärningsmodeller för att predicera arbetskraftstrender inom hälso- och sjukvården i Sverige. Med hjälp av en linjär regressionmodell, en Gradient Boosting Regressor-modell och en Exponential Smoothing-modell syftar forskningen för detta arbete till att ge viktiga insikter för underlaget till makroekonomiska överväganden och att ge en djupare förståelse av Beveridge-kurvan i ett sammanhang relaterat till hälso- och sjukvårdssektorn. Trots vissa utmaningar med datan är målet att förbättra noggrannheten och effektiviteten i beslutsfattandet rörande arbetsmarknaden. Resultaten av denna studie visar maskininlärningspotentialen i predicering i ett ekonomiskt sammanhang, även om inneboende begränsningar och etiska överväganden beaktas. / This study examines the use of machine learning models to predict workforce trends in the healthcare sector in Sweden. Using a Linear Regression model, a Gradient Boosting Regressor model, and an Exponential Smoothing model the research aims to grant needed insight for the basis of macroeconomic considerations and to give a deeper understanding of the Beveridge Curve in the healthcare sector’s context. Despite some challenges with data, the goal is to improve the accuracy and efficiency of the policy-making around the labor market. The results of this study demonstrates the machine learning potential in the forecasting within an economic context, although inherent limitations and ethical considerations are considered.

In recent years, advanced machine learning (ML) models have revolutionized industries ranging from the healthcare sector to retail and E-commerce. However, these models have become increasingly complex, making it difficult for even domain experts to understand and retrace the model's decision-making process. To address this challenge, several frameworks for explainable AI have been proposed and developed. This thesis focuses on counterfactual explanations (CFEs), which provide actionable insights by informing users how to modify inputs to achieve desired outputs. However, computing CFEs for a general black-box ML model is computationally expensive since it hinges on solving a challenging optimization problem. To efficiently solve this optimization problem, we propose using Bayesian optimization (BO), and introduce the novel algorithm Separated Bayesian Optimization (SBO). SBO exploits the formulation of the counterfactual function as a composite function. Additionally, we propose warm-starting SBO, which addresses the computational challenges associated with computing multiple CFEs. By decoupling the generation of a surrogate model for the black-box model and the computation of specific CFEs, warm-starting SBO allows us to reuse previous data and computations, resulting in computational discounts and improved efficiency for large-scale applications. Through numerical experiments, we demonstrate that BO is a viable optimization scheme for computing CFEs for black-box ML models. BO achieves computational efficiency while maintaining good accuracy. SBO improves upon this by requiring fewer evaluations while achieving accuracies comparable to the best conventional optimizer tested. Both BO and SBO exhibit improved capabilities in handling various classes of ML decision models compared to the tested baseline optimizers. Finally, Warm-starting SBO significantly enhances the performance of SBO, reducing function evaluations and errors when computing multiple sequential CFEs. The results indicate a strong potential for large-scale industry applications. / Avancerade maskininlärningsmodeller (ML-modeller) har på senaste åren haft stora framgångar inom flera delar av näringslivet, med allt ifrån hälso- och sjukvårdssektorn till detaljhandel och e-handel. I jämn takt med denna utveckling har det dock även kommit en ökad komplexitet av dessa ML-modeller vilket nu lett till att även domänexperter har svårigheter med att förstå och tolka modellernas beslutsprocesser. För att bemöta detta problem har flertalet förklarbar AI ramverk utvecklats. Denna avhandling fokuserar på kontrafaktuella förklaringar (CFEs). Detta är en förklaringstyp som anger för användaren hur denne bör modifiera sin indata för att uppnå ett visst modellbeslut. För en generell svarta-låda ML-modell är dock beräkningsmässigt kostsamt att beräkna CFEs då det krävs att man löser ett utmanande optimeringsproblem. För att lösa optimeringsproblemet föreslår vi användningen av Bayesiansk Optimering (BO), samt presenterar den nya algoritmen Separated Bayesian Optimization (SBO). SBO utnyttjar kompositionsformuleringen av den kontrafaktuella funktionen. Vidare, utforskar vi beräkningen av flera sekventiella CFEs för vilket vi presenterar varm-startad SBO. Varm-startad SBO lyckas återanvända data samt beräkningar från tidigare CFEs tack vare en separation av surrogat-modellen för svarta-låda ML-modellen och beräkningen av enskilda CFEs. Denna egenskap leder till en minskad beräkningskostnad samt ökad effektivitet för storskaliga tillämpningar.  I de genomförda experimenten visar vi att BO är en lämplig optimeringsmetod för att beräkna CFEs för svarta-låda ML-modeller tack vare en god beräknings effektivitet kombinerat med hög noggrannhet. SBO presterade ännu bättre med i snitt färre funktionsutvärderingar och med fel nivåer jämförbara med den bästa testade konventionella optimeringsmetoden. Både BO och SBO visade på bättre kapacitet att hantera olika klasser av ML-modeller än de andra testade metoderna. Slutligen observerade vi att varm-startad SBO gav ytterligare prestandaökningar med både minskade funktionsutvärderingar och fel när flera CFEs beräknades. Dessa resultat pekar på stor potential för storskaliga tillämpningar inom näringslivet.

<p>The swift emergence and wide-ranging utilization of machine learning (ML) across various industries, including healthcare, transportation, and robotics, have underscored the escalating need for efficient, scalable, and privacy-preserving solutions. Recognizing this, we present an integrated examination of three novel frameworks, each addressing different aspects of distributed learning and privacy issues: Two Timescale Hybrid Federated Learning (TT-HF), Delay-Aware Federated Learning (DFL), and Differential Privacy Hierarchical Federated Learning (DP-HFL). TT-HF introduces a semi-decentralized architecture that combines device-to-server and device-to-device (D2D) communications. Devices execute multiple stochastic gradient descent iterations on their datasets and sporadically synchronize model parameters via D2D communications. A unique adaptive control algorithm optimizes step size, D2D communication rounds, and global aggregation period to minimize network resource utilization and achieve a sublinear convergence rate. TT-HF outperforms conventional FL approaches in terms of model accuracy, energy consumption, and resilience against outages. DFL focuses on enhancing distributed ML training efficiency by accounting for communication delays between edge and cloud. It also uses multiple stochastic gradient descent iterations and periodically consolidates model parameters via edge servers. The adaptive control algorithm for DFL mitigates energy consumption and edge-to-cloud latency, resulting in faster global model convergence, reduced resource consumption, and robustness against delays. Lastly, DP-HFL is introduced to combat privacy vulnerabilities in FL. Merging the benefits of FL and Hierarchical Differential Privacy (HDP), DP-HFL significantly reduces the need for differential privacy noise while maintaining model performance, exhibiting an optimal privacy-performance trade-off. Theoretical analysis under both convex and nonconvex loss functions confirms DP-HFL’s effectiveness regarding convergence speed, privacy performance trade-off, and potential performance enhancement with appropriate network configuration. In sum, the study thoroughly explores TT-HF, DFL, and DP-HFL, and their unique solutions to distributed learning challenges such as efficiency, latency, and privacy concerns. These advanced FL frameworks have considerable potential to further enable effective, efficient, and secure distributed learning.</p>

This thesis addresses the challenge of denoising microscopy images captured under low-light conditionswith varying intensity levels. The study compares three deep learning models — N2V, CARE, andRCAN — against the collaborative filter BM4D, which serves as a reference point. The models weretrained on two distinct datasets: Endoplasmic Reticulum and Mitochondria datasets, both acquired witha lattice light-sheet microscope.Results show that BM4D maintains stable performance metrics and delivers superior visual quality,when compared to the noisy input. In contrast, the deep learning models exhibit poor performance onnoisy test images when trained on datasets with non-uniform noise levels. Additionally, a sensitivitycomparison of neural parameter between the same models was made. Revealing that supervised modelsare data-specific to some extent, whereas the self-supervised N2V demonstrates consistent neuralparameters, suggesting lower data specificity. / Denna uppsats tar upp problemet med att reducera brus i mikroskopibilder tagna under svagaljusförhållanden med varierande intensitetsnivåer. Studien jämför tre djupinlärningsmodeller – N2V,CARE och RCAN – mot det kollaborativa filtret BM4D, vilket agerar som en referenspunkt.Modellerna tränades på två olika dataset: Endoplasmic Reticulum och Mitochondria, båda tagna meden selektiv planbelysningsmikroskop (lattice light-sheet microscope).Resultaten visar att BM4D behåller stabila prestationsmått och levererar bättre visuell kvalitet, jämförtmed den brusiga input. Däremot visar djupinlärningsmodellerna bristande prestanda på brusigatestbilder när de tränats på data med icke-enhetliga brusnivåer. Dessutom gjordes enkänslighetsjämförelse av neurala parametrar mellan samma modeller. Detta visade att de övervakademodellerna är specifika för data i viss utsträckning, medan den självövervakade N2V-modellen visarlika neurala parametrar, vilket tyder på lägre dataspecificitet

<p dir="ltr">State-level policy decisions significantly influence various aspects of our daily lives, such as access to healthcare and education. Despite their importance, there is a limited understanding of how these policies and decisions are formulated within the legislative process. This dissertation aims to bridge that gap by utilizing data-driven methods and the latest advancements in machine learning (ML) and natural language processing (NLP). By leveraging data-driven approaches, we can achieve a more objective and comprehensive understanding of policy formation. The incorporation of ML and NLP techniques aids in processing and interpreting large volumes of complex legislative texts, uncovering patterns and insights that might be overlooked through manual analysis. In this dissertation, we pose new analytical questions about the state legislative process and address them in three stages:</p><p><br></p><p dir="ltr">First, we aim to understand the language of political agreement and disagreement in legislative texts. We introduce a novel NLP/ML task: predicting significant conflicts among legislators and sharp divisions in their votes on state bills, influenced by factors such as gender, rural-urban divides, and ideological differences. To achieve this, we construct a comprehensive dataset from multiple sources, linking state bills with legislators’ information, geographical data about their districts, and details about donations and donors. We then develop a shared relational and textual deep learning model that captures the interactions between the bill’s text and the legislative context in which it is presented. Our experiments demonstrate that incorporating this context enhances prediction accuracy compared to strong text-based models.</p><p><br></p><p dir="ltr">Second, we analyze the impact of legislation on relevant stakeholders, such as teachers in education bills. We introduce this as a new prediction task within our framework to better understand the state legislative process. To address this task, we enhance our modeling and expand our dataset using various techniques, including crowd-sourcing, to generate labeled data. This approach also helps us decode legislators’ decision-making processes and voting patterns. Consequently, we refine our model to predict the winners and losers of bills, using this information to more accurately forecast the legislative body’s vote breakdown based on demographic and ideological criteria.</p><p><br></p><p dir="ltr">Third, we enhance our analysis and modeling of state-level bills and policies using two techniques: We normalize the inconsistent, verbose, and complex language of state policies by leveraging Generative Large Language Models (LLMs). Additionally, we evaluate the policies within a broader network context by expanding the number of US states analyzed from 3 to 50 and incorporating new data sources, such as interest groups’ ratings of legislators and public information on legislators’ positions on various issues.</p><p><br></p><p dir="ltr">By following these steps in this dissertation, we aim to better understand the legislative processes that shape state-level policies and their far-reaching effects on society.</p>