Technological advancements made in recent decades in the fields of artificial intelligence (AI) and machine learning (ML) has lead to further automation of tasks previously performed by humans. Manually reviewing and assessing content uploaded to social media and marketplace platforms is one of said tasks that is both tedious and expensive to perform, and could possibly be automated through ML based systems. When introducing ML model predictions to a human decision making process, interpretability and explainability of models has been proven to be important factors for humans to trust in individual sample predictions. This thesis project aims to explore the performance of interpretable ML models used together with humans in an ad review process for a rental marketplace platform. Utilizing the XGBoost framework and SHAP for interpretable ML, a system was built with the ability to score an individual ad and explain the prediction with human readable sentences based on feature importance. The model reached an ROC AUC score of 0.90 and an Average Precision score of 0.64 on a held out test set. An end user survey was conducted which indicated some trust in the model and an appreciation for the local prediction explanations, but low general impact and helpfulness. While most related work focus on model performance, this thesis contributes with a smaller model usability study which can provide grounds for utilizing interpretable ML software in any manual decision making process.

The rapid digitalization of the healthcare domain in recent years highlighted the need for advanced predictive methods particularly based upon deep learning methods. Deep learning methods which are capable of dealing with time- series data have recently emerged in various fields such as natural language processing, machine translation, and the Intensive Care Unit (ICU). The recent applications of deep learning in ICU have increasingly received attention, and it has shown promising results for different clinical tasks; however, there is still a need for the benchmark models as far as a handful of public datasets are available in ICU. In this thesis, a novel benchmark model of four clinical tasks on a multi-center publicly available dataset is presented; we employed deep learning models to predict clinical studies. We believe this benchmark model can facilitate and accelerate the research in ICU by allowing other researchers to build on top of it. Moreover, we investigated the effectiveness of the proposed method to predict the risk of delirium in the varying observation and prediction windows, the variable ranking is provided to ease the implementation of a screening tool for helping caregivers at the bedside. Ultimately, an attention-based interpretable neural network is proposed to predict the outcome and rank the most influential variables in the model predictions’ outcome. Our experimental findings show the effectiveness of the proposed approaches in improving the application of deep learning models in daily ICU practice.

The commercial deployment of automated vehicles (AVs) is around the corner. With the development of automation technology, automobile and IT companies have started to test automated vehicles. Waymo, an automated driving technology development company, has recently opened the self-driving service to the public. The advancement in this emerging mobility option also drives transportation reasearchers and urban planners to conduct automated vehicle-related research, especially to gain insights on the impact of automated vehicles (AVs) in order to inform policymaking. However, the variation with urban form, the heterogeneity of mode choice, and the impacts at disaggregated levels lead to the dynamics of the impacts of AVs, which not comprehensively understood yet. Therefore, this dissertation extends existing knowledge base by understanding the dynamics of the impacts from three perspectives: (1) examining the role of urban form in the performance of SAV systems; (2) exploring the heterogeneity of AV mode choices across regions; and (3) investigating the distribution of energy consumption in the era of AVs. To examine the first aspect, Shared AV (SAV) systems are simulated for 286 cities and the simulation outcomes are regressed on urban form variables that measure density, diversity, and design. It is suggested that the compact development, a multi-core city pattern, high level of diversity, as well as more pedestrian-oriented networks can promote the performance of SAVs measured using service efficiency, trip pooling success rate, and extra VMT generation. The AV mode choice behaviors of private conventional vehicle (PCV) users in Seattle and Knasas City metropolitan areas are examined using an interpretable machine learning framework based on an AV mode choice survey. It is suggested that attitudes and trip and mode-specific attributes are the most predictive. Positive attitudes can promote the adoption of PAVs. Longer PAV in-vehicle time encourages the residents to keep the PCVs. Longer walking distance promotes the usage of SAVs. In addition, the effects of in-vehicle time and walking distance vary across the two examined regions due to distinct urban form, transportation infrustructure and cultural backgrounds. Kansas City residents can tolerate shorter walking distance before switching to SAV choices due to the car-oriented environment while Seattle residents are more sensitive to in-vehicle travel time because of the local congestion levels. The final part of the dissertation examines the demand for energy of AVs at disaggregated levels incorporating heterogeneity of AV mode choices. A three-step framework is employed including the prediction of mode choice, the determination of vehicle trajectories, and the estimation of the demand for energy. It is suggested that the AV scenario can generate -0.36% to 2.91% extra emissions and consume 2.9% more energy if gasoline is used. The revealed distribution of traffic volume suggests that the demand for charging is concentrated around the downtown areas and on highways if AVs consume electricity. In summary, the dissertation demonstrates that there is a dynamics with regard to the impacts and performance of AVs across regions due to various urban form, infrastructure and cultural environment, and the spatial heterogeneity within cities. / Doctor of Philosophy / Automated vehicles (AVs) have been a hot topic in recent years especially after various IT and automobile companies announced their plans for making AVs. Waymo, an automated driving technology development company, has recently opened the self-driving service to the public. Automated vehicles, which are defined as being able to self-drive, self-park, and automate routing, provide potentials for new business models such as privately owned automated vehicles (PAVs) that serve trips within households, shared AVs (SAVs) that offer door-to-door service to the public who request service using app-based platforms, and SAVs with pool where multiple passengers may be pooled together when the vehicles do not detour much if sequentially picking up and dropping off passengers. Therefore, AVs can transform the transportation system especially by reducing vehicle ownership and increasing travel distance. To plan for a sustainable future, it is important to gain an understanding of the impacts of AVs under various scenarios. Thus, a wealth of case studies explore the system performance of SAVs such as served trips per SAV per day. However, the impacts of AVs are not static and tend to vary across cities, depend on heterogeneous mode choices within regions, and may not be evenly distributed within a city. Therefore, this dissertation fills the research gaps by (1) investigating how urban features such as density may influence the system performance of SAVs; (2) exploring heterogeneity of key factors that influence the decisions about using AVs across regions; and (3) examining the distribution of the demand for energy in the era of AVs. The first study in the dissertation simulates the SAVs that serve trips within 286 cities and examines the relationship between the system performance of SAVs and city features such as density, diversity, and design. The system performance of SAVs is evaluated using served trips per SAV per day, percent of pooled trips that allow ridesharing, and percent of extra Vehicle Miles Traveled (VMT) compared to the VMT requested by the served trips. The results suggest that compact diverse development patterns and pedestrian-oriented networks can promote the performance of SAVs. The second study uses an interpretable machine learning framework to understand the heterogeneous mode choice behaviors of private car users in the era of AVs in two regions. The framework uses an AV mode choice survey, where respondents are asked to take mode choice experiments given attributes about the trips, to train machine learning models. Accumulated Local Effects (ALE) plots are used to analyze the model results. ALE outputs the accumulated change of the probability of choosing specific modes within small intervals across the range of the variable of interest. It is suggested that attitudes and trip-specific attributes such as in-vehicle time are the most important determinants. Positive attitudes, longer trips, and longer walking distance can promote the adoption of AV modes. In addition, the effects of in-vehicle time and walking distance vary across the two examined regions due to distinct urban form, transportation infrastructure, and cultural backgrounds. Kansas City residents can tolerate shorter walking distance before switching to SAV choices due to the car-oriented environment while Seattle residents are more sensitive to in-vehicle travel time because of the local congestion levels. The final part of the dissertation examines the demand for energy of AVs at disaggregated levels incorporating heterogeneity of AV mode choices. A three-step framework is employed including the prediction of mode choice, the determination of vehicle trajectories, and the estimation of the demand for energy. It is suggested that the AV scenario can generate -0.36% to 2.91% of extra emissions and consume 2.9% more energy compared to a business as usual (BAU) scenario if gasoline is used. The revealed distribution of traffic volume suggests that the demand for charging is concentrated around the downtown areas and on highways if AVs consume electricity. In summary, the dissertation demonstrates that there is a dynamics with regard to the impacts and performance of AVs across regions due to various urban form, infrastructure and cultural environment, and the spatial heterogeneity within cities.

The use of clinical decision support (CDS) systems has grown progressively during the past decades. CDS systems are associated with improved patient safety and outcomes, better prescription and diagnosing practices by clinicians and lower healthcare costs. Quality assurance of these systems is critical, given the potentially severe consequences of any errors. Yet, after several decades of research, there is still no consensual or standardized approach to their verification and validation (V&V). This project is a descriptive and exploratory case study aiming to provide a practical description of how Cambio CDS, a market-leading developer of CDS services, conducts its V&V process. Qualitative methods including semi-structured interviews and coding-based textual data analysis were used to elicit the description of the V&V approaches used by the company. The results showed that the company’s V&V methodology is strongly influenced by the company’s model-driven development approach, a strong focus and leveraging of domain knowledge and good testing practices with a focus on automation and test-driven development. A few suggestions for future directions were discussed.

We integrate theoretical, computational and data-driven approaches for the sake of understanding, design and discovery of metal based catalysts. Firstly, we develop theoretical frameworks for predicting electronic descriptors of transition and noble metal alloys, including a physics model of d-band center, and a tight-binding theory of d-band moments to systematically elucidate the distinct electronic structures of novel catalysts. Within this framework, the hybridization of semi-empirical theories with graph neural network and attribution analysis enables accurate prediction equipped with mechanistic insights. In addition, novel physics effect controlling surface reactivity beyond conventional understanding is uncovered. Secondly, we develop a computational and data-driven framework to model high entropy alloy (HEA) catalysis, incorporating thermodynamic descriptor-based phase stability evaluation, surface segregation modeling by deep learning potential-driven molecular simulation and activity prediction through machine learning-embedded electrokinetic model. With this framework, we successfully elucidate the experimentally observed improved activity of PtPdCuNiCo HEA in oxygen reduction reaction. Thirdly, a Bayesian optimization framework is employed to optimize racemic lactide polymerization by searching for stereoselective aluminum (Al) -complex catalysts. We identified multiple new Al-complex molecules that catalyzed either isoselective or heteroselective polymerization. In addition, feature attribution analysis uncovered mechanistically meaningful ligand descriptors that can access quantitative and predictive models for catalyst development. / Doctor of Philosophy / In addressing the critical issues of climate change, energy scarcity, and pollution, the drive towards a sustainable economy has made catalysis a key area of focus. Computational chemistry has revolutionized our understanding of catalysts, especially in identifying and analyzing their active sites. Furthermore, the integration of open-access data and advanced computing has elevated data science as a crucial component in catalysis research. This synergy of computational and data-driven approaches is advancing the development of innovative catalytic materials, marking a significant stride in tackling environmental challenges. In my PhD research, I mainly work on the development of computational and data-driven methods for better understanding, design and discovery of catalytic materials. Firstly, I develop physics models for people to intuitively understand the reactivity of transition and noble metal catalysts. Then I embed the physics models into deep learning models for accurate and insightful predictions. Secondly, for a class of complex metal catalysts called high-entropy alloy (HEA) which is hard to model, I develop a modeling framework by hybridizing computational and data-driven approaches. With this framework, we successfully elucidate the experimentally observed improved activity of PtPdCuNiCo HEA in oxygen reduction reaction which is a key reaction in fuel cell technology. Thirdly, I develop a framework to virtually screen catalyst molecules to optimize polymerization reaction and provide potential candidates to our experimental collaborator to synthesize. Our collaboration leads to the discovery of novel high-performance molecular catalysts.

Can Machines Explain Stock Returns? Thesis Abstract Karolína Chalupová January 5, 2021 Recent research shows that neural networks predict stock returns better than any other model. The networks' mathematically complicated nature is both their advantage, enabling to uncover complex patterns, and their curse, making them less readily interpretable, which obscures their strengths and weaknesses and complicates their usage. This thesis is one of the first attempts at overcoming this curse in the domain of stock returns prediction. Using some of the recently developed machine learning interpretability methods, it explains the networks' superior return forecasts. This gives new answers to the long- standing question of which variables explain differences in stock returns and clarifies the unparalleled ability of networks to identify future winners and losers among the stocks in the market. Building on 50 years of asset pricing research, this thesis is likely the first to uncover whether neural networks support the economic mechanisms proposed by the literature. To a finance practitioner, the thesis offers the transparency of decomposing any prediction into its drivers, while maintaining a state-of-the-art profitability in terms of Sharpe ratio. Additionally, a novel metric is proposed that is particularly suited...

In the past few years, Artificial Intelligence (AI) has evolved into a powerful tool applied in multi-disciplinary fields to resolve sophisticated problems. As AI becomes more powerful and ubiquitous, oftentimes the AI methods also become opaque, which might lead to trust issues for the users of the AI systems as well as fail to meet the legal requirements of AI transparency. In this report, the possibility of making a credit-card fraud detection support system explainable to users is investigated through a quantitative survey. A publicly available credit card dataset was used. Deep Learning and Random Forest were the two Machine Learning (ML) methodsimplemented and applied on the credit card fraud dataset, and the performance of their results was evaluated in terms of their accuracy, recall, sufficiency, and F1 score. After that, two explainable AI (XAI) methods - SHAP (Shapley Additive Explanations) and LIME (Local Interpretable Model-Agnostic Explanations) were implemented and applied to the results obtained from these two ML methods. Finally, the XAI results were evaluated through a quantitative survey. The results from the survey revealed that the XAI explanations can slightly increase the users' impression of the system's ability to reason and LIME had a slight advantage over SHAP in terms of explainability. Further investigation of visualizing data pre-processing and the training process is suggested to offer deep explanations for users.

As more data is being gathered in industrial production facilities, the interest in applying machine learning models to the data is growing. This includes the iron ore mining industry, and in particular the build-up of slag in grate-kiln furnaces. Slag is a byproduct in the pelletizing process within these furnaces, that can cause production stops, quality issues, and unplanned maintenance. Previous studies on slag build-up have been done mainly by chemists and process engineers. Whilst previous research has hypothesized contributing factors to slag build-up, the studies have mostly been conducted in simulation environments and thus have not used real sensor data utilizing machine learning models. Luossavaara-Kiirunavaara Aktiebolag (LKAB) has provided data from one of their grate-kiln furnaces, a time-series data of sensor readings, that compressed before storage.  A Scala package was built to ingest and interpolate the LKAB data and make it ready for machine learning experiments. The estimation of slag within the kiln was found too arbitrary to make accurate predictions. Therefore, three quality metrics, tightly connected to the build-up of slag, were selected as target variables instead. Independent and identically distributed (IID) units of data were created by isolating fuel usage, product type produced and production rate. Further, another IID criterion was created, adjusting the time for each feature in order to be able to compare feature values for a single pellet in production. Specifically, the time it takes for a pellet to go from the feature sensor to the quality test was added to the original timestamp. This resulted in a table where each row represents multiple features and quality measures for the same small batch of pellets. An IID unit of interest was then used to find the most contributing features by using principal component analysis (PCA) and lasso regression. It was found that using the two mentioned methods, the number of features could be reduced to a smaller set of important features. Further, using decision tree regression with the subset of features, selected from the most important features, it was found that decision tree regression had a similar performance with the subset of features as the lasso regression. Decision tree and lasso regression were chosen for interpretability, which was important in order to be able to discuss the contributing factors with LKAB process engineers. / Idag genereras allt mer data från industriella produktionsanläggningar och intresset att applicera maskininlärningsmodeller på denna data växer. Detta inkluderar även industrin för utvining av järnmalm, i synnerhet uppbyggnaden av slagg i grate-kiln ugnar. Slagg är en biprodukt från pelletsproduktionen som kan orsaka produktionsstopp, kvalitetsbrister och oplanerat underhåll av ugnarna. Tidigare forskning kring slagguppbyggnad har i huvudsak gjorts av kemister och processingenjörer och ett antal bidragande faktorer till slagguppbyggnad ha antagits. Däremot har dessa studier främst utförts i simulerad experimentmiljö och därför inte applicerat maskininlärningsmodeler på sensordata från produktion. Luossavaara-Kiirunavaara Aktiebolag (LKAB) har till denna studie framställt och försett data från en av deras grate-kiln ugnar, specifikt tidsseriedata från sensorer som har komprimerats innan lagring. Ett Scala-paket byggdes för att ladda in och interpolera LKAB:s data, för att sedan göra den redo och applicerbar för experiment med maskininlärningsmodeller. Direkta mätningar för slagguppbyggnad och slaggnivå upptäcktes vara för slumpartade och bristfälliga för prediktion, därför användas istället tre kvalitetsmätningar, med tydligt samband till påföljderna från slagguppbyggnad, som målvariabler. Independent and identically distributed (IID) enheter skapades för all data genom att isolera bränsleanvändning, produkttyp och produktionstakt. Vidare, skapades ytterligare ett kriterie för IID:er, en tidsjustering av varje variabel för att göra det möjligt att kunna jämföra variabler inbördes för en enskild pellet i produktion. Specifikt, användes tiden det tar för en pellet från att den mäts av en enskild sensor till att kvalitetstestet tas. Tidsskillnaden adderas sedan till sensormätningens tidsstämpel. Detta resulterade i en tabell där varje rad representerade samma lilla mängd av pellets. En IID enhet av intresse analyserades sedan för att undersöka vilka variabler som har störst varians och påverkan genom en principal komponentsanalys (PCA) och lassoregression. Genom att använda dessa metoder konstaterades det att antalet variabler kunde reduceras till ett mindre antal variabler och ett nytt, mindre, dataset av de viktigaste variablerna skapades. Vidare, genom regression av beslutsträd med de viktigaste variablerna, konstaterades att beslutträdsregression och lassoregression hade liknande prestanda när data med de viktigaste variablerna användes. Beslutträdsregression och lassoregression användes för att experimentens resultat skulle ha en hög förklaringsgrad, vilket är viktigt för att kunna diskutera variabler med högst påverkan på slagguppbyggnaden och ge resultat som är tolkbara och användbara för LKAB:s processingenjörer.

Sympathovagal imbalance is known to precede the on-set of atrial fibrillation (AF) and has been analyzed extensively based on heart rate variability (HRV). However, the relationship between sympathetic and vagal effects before AF onset and their influence on various HRV features have not been fully elucidated. QT interval variability (QTV) reflects sympathetic activity and may therefore provide further insights into this relationship. Using the time delay stability (TDS) method, we investigated temporal changes in coupling behavior before AF onset between 20 vagal or sympathovagal-associated HRV and QTV features. We applied the TDS method to 26 electrocardiograms from the MIT-BIH AF database with at least one hour of sinus rhythm preceding AF onset. Sinus rhythm segments were split into 5-minute windows with 50 % overlap. Logistic regression analysis revealed significantly (p<0.01) increased coupling between QTV and vagal HRV features from 20 to 15 minutes before AF onset. We found similar behavior between QTV and sympathovagal HRV features. This indicates sympathetic predominance increasing until 15 minutes before the onset of AF and decreasing towards vagal predominance right before AF onset. Our results provide new insights into temporal changes of sympathovagal imbalance preceding AF onset and may improve the prediction of AF in clinical applications.

This Master's Thesis project set out with the objective to propose a machine learning model for predicting insurance risk at the level of an individual coverage, and compare it towards the existing models used by the project provider Gjensidige Försäkring. Due to interpretability constraints, it was found that this problem can be translated into a standard tabular regression task, with well defined target distributions. However, it was early identified that the set of feasible models do not contain pure black box models such as XGBoost, LightGBM and CatBoost which are typical choices for tabular data regression. In the report, we explicitly formulate the interpretability constraints in sharp mathematical language. It is concluded that interpretability can be ensured by enforcing a particular structure on the Hilbert space across which we are looking for the model.  Using this formalism, we consider two different approaches for fitting high performing models that maintain interpretability, where we conclude that gradient boosted regression tree based Generalized Additive Models in general, and the Explainable Boosting Machine in particular, is a promising model candidate consisting of functions within the Hilbert space of interest. The other approach considered is the basis expansion approach, which is currently used at the project provider. We make the argument that the gradient boosted regression tree approach used by the Explainable Boosting Machine is a more suitable model type for an automated, data driven modelling approach which is likely to generalize well outside of the training set. Finally, we perform an empirical study on three different internal datasets, where the Explainable Boosting Machine is compared towards the current production models. We find that the Explainable Boosting Machine systematically outperforms the current models on unseen test data. There are many potential ways to explain this, but the main hypothesis brought forward in the report is that the sequential model fitting procedure allowed by the regression tree approach allows us to effectively explore a larger portion of the Hilbert space which contains all permitted models in comparison to the basis expansion approach. / Detta mastersexamensarbete utgår från målsättningen att föreslå en maskinlärningsmodell för att förutspå försäkringsrisk, på nivån av enskilda försäkringar. Denna modell ska sedan jämföras mot nuvarande modeller som används hos Gjensidige Försäkring, som tillhandahåller projektet. Detta problem kan formuleras som ett traditionellt regressionsproblem på tabulär data, med väldefinerade målfördelningar. På grund av begränsningar kring krav på modellens förståbarhet identifierades det tidigt i projektet att mängden av tillåtna modeller inte innehåller ren black box modeller som XGBoost, LightGBM eller CatBoost, vilket är typiska förstahandsval för den här problemklassen. I rapporten formulerar vi förståbarhetskraven i skarpt, matematiskt språk, och drar slutsatsen att önskad förståbarhet kan uppnås genom en specifik struktur på det Hilbertrum där vi letar efter den optimala modellen. Utifrån denna formalism evaluerar vi två olika metoder för att anpassa modeller med god prestanda som uppnår önskade förståbarhetskrav. Vi drar slutsatsen att Generalized Additive Models anpassade till datat genom gradientboostade regressionsträd i allmänhet, och Explainable Boosting Machine i synnerhet är en lovande modellkandidat bestående av funktioner i vårt Hilbertrum av intresse. Vi utvärderar dessutom ett tillvägagångssätt för att anpassa Generalized Additive Models till datat genom basexpansioner, vilket är den metod som primärt används idag hos Gjensidige Försäkring. Vi argumenterar för att metoder som bygger på gradientboostade regressionsträd, såsom Explainable Boosting Machine, är mer lämplig för ett automatiserbart, datadrivet arbetssätt till att bygga modeller som generaliserar väl utanför träningsdatat.  Slutligen genomför vi en empirisk studie på tre olika interna dataset, där Explainable Boosting Machine jämförs mot nuvarande produktionsmodeller, vilka bygger på den tidigare nämnda basexpansionsmetodiken. Vi finner att Explainable Boosting Machine systematiskt överpresterar kontra nuvarande modeller på osedd testdata. Det finns många potentiella förklaringar till detta, men den huvudsakliga hypotsen som diskuteras i denna rapport är att den gradientboostade regressionsträdsmetodiken gör det möjligt att effektivt utforska en större delmängd av det Hilbertrum som innehåller alla tillåtna modeller i jämförelse med basexpansionsmetodiken.