The development of new solutions incorporating artificial intelligence (AI) within the medical field is an area of great interest. However, access to comprehensive and diverse datasets is restricted due to the sensitive nature of the data. A potential solution to this is to generatesynthetic datasets based on real medical data. Synthetic data could protect the integrity of the subjects while preserving the inherent information necessary for training AI models and be generated in greater quantity than otherwise available. This thesis project aims to generate reliable clinical trial subject data using a generative adversarial network (GAN). The main data set used is a mock clinical trial dataset consisting of multiple subject visits, however an additional data set containing authentic medical data is also used for better insights into the model’s ability to learn underlying relationships. The thesis also investigates training strategies for simulating the temporal dimension and the missing values in the data. The GAN model used is an altered version of the Conditional Tabular GAN (CTGAN)made to be compatible with the preprocessed clinical trial mock data, and multiple model architectures and number of training epochs are examined. The results show great potential for GAN models on clinical trial datasets, especially for real-life data. One model, trained on the authentic dataset, generates near-perfect synthetic data with respect to column distributions and correlation between columns. The results also show that classification models trained on synthetic data and tested on real data have the potential to match the performance of classification models trained on real data. While the synthetic data replicates the missing values, no definitive conclusion can be drawn regarding the temporal characteristics due to the sparsity of the mock dataset and lack of real correlations in it. Although the results are promising, further experiments on authentic datasets with less sparsity are required.

The global use of solar power is growing at an unprecedented rate, making the need toaccurately track the energy generation of decentralised solar energy systems (SES) more andmore relevant. The purpose of this thesis is to further develop a binary image classifier for thesimulation system framework known as Alfrödull, which will be used to detect and segment SESfrom aerial images to simulate the energy generation within a given Swedish municipality on anhourly basis. This project focuses on improving the Alfrödull classifier through four differentanalyses. the first focusing on examining how additional training data from publicly availabledatasets affects the model performance. The second on how the model can be improvedthrough the use of various image pre-processing techniques. The third on how the model canbe improved through balancing the training datasets to make up for the low amount of positiveimages as well as utilising model ensembles for joint classification. Finally, the fourth analysisemploys a sliding window approach to classify overlapping image tiles. The results show thathaving training data that is a good representation of the environment the model will be used in iscrucial, that the use of image augmentation policies can significantly improve modelperformance, that compensating for class imbalance as well as utilising ensemble methodspositively impacts model performance and that a sliding window approach to classifyingoverlapping images significantly decreases the amount of missed SES at the cost of clusters offalsely classified negative images (false positives). In conclusion, this thesis serves as animportant stepping stone in the practical implementation of the Alfrödull framework, showcasingthe key aspects in making a well performing binary image classifier of SES in Sweden.

In this thesis the major objective is the implementation of the inverse neural network concept in the design of printed lens (transmitarray) antenna. As it is computationally extensive to perform full-wave simulations for entire transmitarray structure and thereafter perform optimization, the idea is to generate a design database assuming that a unit cell of the transmitarray is situated inside a 2D infinite periodic structure. This way we generate a design database of transmission coefficient by varying the unit cell parameters. Since, for the actual design, we need dimensions for each cell on the transmitarray aperture and to do this we need to invert the design database. The major contribution of this thesis is the proposal and the implementation of database inversion methodology namely inverse neural network modelling. We provide the algorithms for carrying out the inversion process as well as provide check results to demonstrate the reliability of the proposed methodology. Finally, we apply this approach to design a transmitarray antenna, and measure its performance.

The field of autonomous driving is as active as it has ever been, but the reality where an autonomous vehicle can drive on all roads is currently decades away. Instead, using an on-the-fly learning method, such as qHebb learning, a system can,after some demonstration, learn the appearance of any road and take over the steering wheel. By training in a simulator, the amount and variation of training can increase substantially, however, an on-rails auto-pilot does not sufficiently populate the learning space of such a model. This study aims to explore concepts that can increase the variance in the training data whilst the vehicle trains online. Three computationally light concepts are proposed that each manages to result in a model that can navigate through a simple environment, thus performing better than a model trained solely on the auto-pilot. The most noteworthy approach uses multiple thresholds to detect when the vehicle deviates too much and replicates the action of a human correcting its trajectory. After training on less than 300 frames, a vehicle successfully completed the full test environment using this method. / Autonom körning är ett aktivt område inom både industrin och forskarvärlden, men ännu är en verklighet där förarlösa fordon kan ta sig fram oavsett väg, decennier bort. Istället kan man genom att använda en adaptiv inlärningsmodell som qHebb learning uppnå ett system som kan ta sig fram självmant på alla vägar, efter en initial inlärningsperiod. Genom att använda en simulator skulle möjligheten att träna en sådan modell öka avsevärt, likaså variationen av vägtyper och det omgivande landskapet. Dock klarar inte en enformig autopilot att fylla modellens lärningsrymd. Detta arbete stävar efter att utforska koncept som kan öka variationen på träningsdatan, medan fordonet kör. Tre prestandalätta metoder presenteras som alla överträffar autopiloten och resulterar i en modell som lärt sig att följa en väg längs kurvor och raksträckor. Det främsta konceptet använder sig av två tröskelvärden för att korrigera fordonets styrning då den avviker för mycket från den korrekta rutten. Efter träning på färre än 300 bilder lyckas denna metod slutföra alla testsegment utan kollision.