Global ETD Search

91	Multi-fidelity Machine Learning for Perovskite Band Gap Predictions Panayotis Thalis Manganaris (16384500) 16 June 2023 (has links) <p>A wide range of optoelectronic applications demand semiconductors optimized for purpose.</p> <p>My research focused on data-driven identification of ABX3 Halide perovskite compositions for optimum photovoltaic absorption in solar cells.</p> <p>I trained machine learning models on previously reported datasets of halide perovskite band gaps based on first principles computations performed at different fidelities.</p> <p>Using these, I identified mixtures of candidate constituents at the A, B or X sites of the perovskite supercell which leveraged how mixed perovskite band gaps deviate from the linear interpolations predicted by Vegard's law of mixing to obtain a selection of stable perovskites with band gaps in the ideal range of 1 to 2 eV for visible light spectrum absorption.</p> <p>These models predict the perovskite band gap using the composition and inherent elemental properties as descriptors.</p> <p>This enables accurate, high fidelity prediction and screening of the much larger chemical space from which the data samples were drawn.</p> <p><br></p> <p>I utilized a recently published density functional theory (DFT) dataset of more than 1300 perovskite band gaps from four different levels of theory, added to an experimental perovskite band gap dataset of \textasciitilde{}100 points, to train random forest regression (RFR), Gaussian process regression (GPR), and Sure Independence Screening and Sparsifying Operator (SISSO) regression models, with data fidelity added as one-hot encoded features.</p> <p>I found that RFR yields the best model with a band gap root mean square error of 0.12 eV on the total dataset and 0.15 eV on the experimental points.</p> <p>SISSO provided compound features and functions for direct prediction of band gap, but errors were larger than from RFR and GPR.</p> <p>Additional insights gained from Pearson correlation and Shapley additive explanation (SHAP) analysis of learned descriptors suggest the RFR models performed best because of (a) their focus on identifying and capturing relevant feature interactions and (b) their flexibility to represent nonlinear relationships between such interactions and the band gap.</p> <p>The best model was deployed for predicting experimental band gap of 37785 hypothetical compounds.</p> <p>Based on this, we identified 1251 stable compounds with band gap predicted to be between 1 and 2 eV at experimental accuracy, successfully narrowing the candidates to about 3% of the screened compositions.</p> Compound semiconductors Organic semiconductors Data engineering and data science halide perovskites band gap feature extraction and representation SISSO random forest regression analysis Gaussian Process Regression Analysis SHapley Additive exPlanations (SHAP) combinatorial datasets data augmentation method Lead Free Perovskite Solar Cells Density Functional Theory (DFT) multi-fidelity data multi-task learning (MTL)
92	Computationally Efficient Explainable AI: Bayesian Optimization for Computing Multiple Counterfactual Explanantions / Beräkningsmässigt Effektiv Förklarbar AI: Bayesiansk Optimering för Beräkning av Flera Motfaktiska Förklaringar Sacchi, Giorgio January 2023 (has links) 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. Explainable AI Counterfactual Explanations (CFEs) Bayesian Optimization (BO) Black-Box Models Model-Agnostic Machine Learning (ML) Efficient Computation High-Stake Decisions Förklarbar AI Kontrafaktuell Förklaring (CFE) Bayesiansk Optimering (BO) Svarta lådmodeller Modellagnostisk Maskininlärning Beräkningsmässigt Effektiv Beslut med höga insatser Other Mathematics Annan matematik
93	Die voedselparadoks : 'n ondersoek na vraagstukke rondom voedselsekuriteit in Suid-Afrika Kotzé, Derica Alba 11 1900 (has links) Text in Afrikaans / Summaries in Afrikaans and English / Miljoene mense ervaar voedselonsekerheid en een uit elke 50 hanger mense is woonagtig in Suid Afrika. Daar is genoeg voedsel op ons planeet om elke mens van 'n voldoende voorraad voedsel te verseker; dit waarborg egter nie voedselsekuriteit aan almal nie. Dit is die voedselparadoks: ondanks globale surplusproduksie van voedsel, ly miljoene mense wereldwyd aan wanvoeding en honger, maar veral in die ontwikkelende lande. Suid-Afrika is geen uitsondering nie en ten spyte van selfvoorsiening in voedsel, balanseer die voedselgelykstelling nie. Daar bestaan 'n ekstreme gaping tussen die produksie en verbruik van voedsel. Gevolglik is die probleem wat nagevors is in hierdie studie die gebrek aan voedselsekuriteit binne 'n wereldkonteks met voedselsurplusse en hoe dit reflekteer in Suid-Afrika. Teen hierdie agtergrond is daar 'n studie gedoen van die oorsake van voedselonsekerheid en die teoriee en verduidelikings van hongersnood. Die fokus van hierdie navorsingstudie is drieledig van aard. Eerstens fokus dit op 'n konseptuele ondersoek na hanger, armoede, voedselsekuriteit en hongersnood in Afrika. Tweedens is ondersoek ingestel na die oorsake vir die gebrek aan voedselsekuriteit in Afrika. Derdens is daar gefokus op Suid-Afrika en is 'n ondersoek gedoen na die voorkoms van hanger, wanvoeding, armoede en die nasionale konteks van voedselsekuriteit met die doel om vraagstukke daaromheen te identifiseer. Daar is bevind dat voedselsekuriteit bepaal word deur die beskikbaarheid van voedsel (aanbod) en die vermoe van mense om dit te bekom (aanvraag). Dit blyk dat die ontwikkelingsproses, regeringsbeleid, ekologiese omgewing en tegnologie, wetenskap en navorsing 'n direkte invloed het op die voedselsekuriteit van mense, en dat Suid-Afrika nie verskil van ander Afrikalande in hierdie verband nie. Hoewel Suid-Afrika voedselselfvoorsiening bereik het, ly miljoene mense honger weens armoede en die gebrek aan aansprake wat bydra tot 'n gebrek aan voedselsekuriteit. Die studie toon dat die Suid-Afrikaanse regering verskeie beleidsmaatreels in plek het ter bevordering van voedselsekuriteit, maar dat dit nie in die praktyk verwesenlik word nie. / Millions of people in the world experience food insecurity and one out ofevery 50 hungry people lives in South Africa. There is enough food on our planet to assure every person of an adequate supply of food; however, this does not guarantee food security for all. This is the food paradox: despite a global surplus production of food, millions of people experience malnutrition and hunger all over the world, but especially in the developing countries. South Africa is no exception and despite self-sufficiency in food, the food equation is not balanced. An extreme gap exists between the production and consumption of food. Consequently, the problem researched in this study is the lack of food security in a world context with surplus food and how this is reflected in South Africa. Against this background a study was undertaken of the causes of food insecurity and the theories and explanations of famine. The focus of this research study is threefold. Firstly it focuses on a conceptual enquiry intohunger, poverty, food security and famine in Africa. Secondly there is an enquiry into the causes of the lack of food security in Africa. Thirdly it focuses on South Africa and an enquiry is done into the incidence of hunger, malnutrition and poverty, and into the national context of food security with the aim of identifying relevant problems in food security. It was found that food security is determined by the availability of food (supply) and the capability of people to obtain it (demand). It appears that the development process, government policy, ecological environment and technology, science and research directly affect the food security of people, and that South Africa does not differ from other African countries in this regard. Although South Africa has achieved food self-sufficiency, millions of people experience hunger because of poverty and the lack of entitlements. The study shows that the South African government has various policy measures for the promotion of food security in place, but that food security does not materialise in practice. / Development Studies / D.Litt. et Phil. (Ontwikkelingsadministrasie) Food security Food self-sufficiency Food paradox Theories of famine National context of food security Urban bias Export agriculture Cash crops Consumption Trade Price policy food markets Agricultural policy Commercial agriculture Subsistence agriculture Food policy 363.80968 Food industry and trade -- South Africa. Food supply -- Africa. Food supply -- South Africa. Agriculture and state -- Africa. Agriculture and state -- South Africa. Agriculture -- Africa. Agriculture -- South Africa. Famines -- Africa. Famines -- South Africa. Agricultural productivity -- Africa. Africa -- Economic conditions. South Africa -- Economic conditions. Africa -- Economic policy. South Africa -- Economic policy.
94	Die voedselparadoks : 'n ondersoek na vraagstukke rondom voedselsekuriteit in Suid-Afrika Kotzé, Derica Alba 11 1900 (has links) Text in Afrikaans / Summaries in Afrikaans and English / Miljoene mense ervaar voedselonsekerheid en een uit elke 50 hanger mense is woonagtig in Suid Afrika. Daar is genoeg voedsel op ons planeet om elke mens van 'n voldoende voorraad voedsel te verseker; dit waarborg egter nie voedselsekuriteit aan almal nie. Dit is die voedselparadoks: ondanks globale surplusproduksie van voedsel, ly miljoene mense wereldwyd aan wanvoeding en honger, maar veral in die ontwikkelende lande. Suid-Afrika is geen uitsondering nie en ten spyte van selfvoorsiening in voedsel, balanseer die voedselgelykstelling nie. Daar bestaan 'n ekstreme gaping tussen die produksie en verbruik van voedsel. Gevolglik is die probleem wat nagevors is in hierdie studie die gebrek aan voedselsekuriteit binne 'n wereldkonteks met voedselsurplusse en hoe dit reflekteer in Suid-Afrika. Teen hierdie agtergrond is daar 'n studie gedoen van die oorsake van voedselonsekerheid en die teoriee en verduidelikings van hongersnood. Die fokus van hierdie navorsingstudie is drieledig van aard. Eerstens fokus dit op 'n konseptuele ondersoek na hanger, armoede, voedselsekuriteit en hongersnood in Afrika. Tweedens is ondersoek ingestel na die oorsake vir die gebrek aan voedselsekuriteit in Afrika. Derdens is daar gefokus op Suid-Afrika en is 'n ondersoek gedoen na die voorkoms van hanger, wanvoeding, armoede en die nasionale konteks van voedselsekuriteit met die doel om vraagstukke daaromheen te identifiseer. Daar is bevind dat voedselsekuriteit bepaal word deur die beskikbaarheid van voedsel (aanbod) en die vermoe van mense om dit te bekom (aanvraag). Dit blyk dat die ontwikkelingsproses, regeringsbeleid, ekologiese omgewing en tegnologie, wetenskap en navorsing 'n direkte invloed het op die voedselsekuriteit van mense, en dat Suid-Afrika nie verskil van ander Afrikalande in hierdie verband nie. Hoewel Suid-Afrika voedselselfvoorsiening bereik het, ly miljoene mense honger weens armoede en die gebrek aan aansprake wat bydra tot 'n gebrek aan voedselsekuriteit. Die studie toon dat die Suid-Afrikaanse regering verskeie beleidsmaatreels in plek het ter bevordering van voedselsekuriteit, maar dat dit nie in die praktyk verwesenlik word nie. / Millions of people in the world experience food insecurity and one out ofevery 50 hungry people lives in South Africa. There is enough food on our planet to assure every person of an adequate supply of food; however, this does not guarantee food security for all. This is the food paradox: despite a global surplus production of food, millions of people experience malnutrition and hunger all over the world, but especially in the developing countries. South Africa is no exception and despite self-sufficiency in food, the food equation is not balanced. An extreme gap exists between the production and consumption of food. Consequently, the problem researched in this study is the lack of food security in a world context with surplus food and how this is reflected in South Africa. Against this background a study was undertaken of the causes of food insecurity and the theories and explanations of famine. The focus of this research study is threefold. Firstly it focuses on a conceptual enquiry intohunger, poverty, food security and famine in Africa. Secondly there is an enquiry into the causes of the lack of food security in Africa. Thirdly it focuses on South Africa and an enquiry is done into the incidence of hunger, malnutrition and poverty, and into the national context of food security with the aim of identifying relevant problems in food security. It was found that food security is determined by the availability of food (supply) and the capability of people to obtain it (demand). It appears that the development process, government policy, ecological environment and technology, science and research directly affect the food security of people, and that South Africa does not differ from other African countries in this regard. Although South Africa has achieved food self-sufficiency, millions of people experience hunger because of poverty and the lack of entitlements. The study shows that the South African government has various policy measures for the promotion of food security in place, but that food security does not materialise in practice. / Development Studies / D.Litt. et Phil. (Ontwikkelingsadministrasie) Food security Food self-sufficiency Food paradox Theories of famine National context of food security Urban bias Export agriculture Cash crops Consumption Trade Price policy food markets Agricultural policy Commercial agriculture Subsistence agriculture Food policy 363.80968 Food industry and trade -- South Africa. Food supply -- Africa. Food supply -- South Africa. Agriculture and state -- Africa. Agriculture and state -- South Africa. Agriculture -- Africa. Agriculture -- South Africa. Famines -- Africa. Famines -- South Africa. Agricultural productivity -- Africa. Africa -- Economic conditions. South Africa -- Economic conditions. Africa -- Economic policy. South Africa -- Economic policy.
95	The Darwinian revolution as a knowledge reorganization / a historical-epistemological analysis and a reception analysis based on a novel model of scientific theories Zacharias, Sebastian 24 February 2015 (has links) Die Dissertation leistet drei Beiträge zur Forschung: (1) Sie entwickelt ein neuartiges vierstufiges Modell wissenschaftlicher Theorien. Dieses Modell kombiniert logisch-empiristische Ansätze (Carnap, Popper, Frege) mit Konzepten von Metaphern & Narrativen (Wittgenstein, Burke, Morgan), erlaubt so deutlich präzisiere Beschreibungen wissenschaftlicher Theorien bereit und löst/mildert Widersprüche in logisch-empiristischen Modellen. (Realismus vs. Empirismus, analytische vs. synthetische Aussagen, Unterdeterminiertheit/ Holismus, wissenschaftliche Erklärungen, Demarkation) (2) Mit diesem Modell gelingt ein Reihenvergleich sechs biologischer Theorien von Lamarck (1809), über Cuvier (1811), Geoffroy St. Hilaire (1835), Chambers (1844-60), Owen (1848-68), Wallace (1855/8) zu Darwin (1859-1872). Dieser Vergleich offenbart eine interessante Asymmetrie: Vergleicht man Darwin mit je einem Vorgänger, so bestehen zahlreiche wichtige Unterschiede. Vergleicht man ihn mit fünf Vorgängern, verschwinden diese fast völlig: Darwins originärer Beitrag zur Revolution in der Biologie des 19.Jh ist klein und seine Antwort nur eine aus einer kontinuierlichen Serie auf die empirischen Herausforderungen durch Paläontologie & Biogeographie seit Ende des 18. Jh. (3) Eine gestufte Rezeptionsanalyse zeigt, warum wir dennoch von einer Darwinschen Revolution sprechen. Zuerst zeigt eine quantitative Analyse der fast 2.000 biologischen Artikel in Britannien zwischen 1858 und 1876, dass Darwinsche Konzepte zwar wichtige Neuerungen brachten, jedoch nicht singulär herausragen. Verlässt man die Biologie und schaut sich die Rezeption bei anderen Wissenschaftlern und gebildeten Laien an, wechselt das Bild: Je weiter man aus der Biologie heraustritt, desto weniger Ebenen biologischen Wissens kennen die Rezipienten und desto sichtbarer wird Darwins Beitrag. Schließlich findet sich sein Beitrag in den abstraktesten Ebenen des biologischen Wissens: in Narrativ und Weltbild – den Ebenen die Laien rezipieren. / The dissertation makes three contributions to research: (1) It develops a novel 4-level-model of scientific theories which combines logical-empirical ideas (Carnap, Popper, Frege) with concepts of metaphors & narratives (Wittgenstein, Burke, Morgan), providing a new powerful toolbox for the analysis & comparison of scientific theories and overcoming/softening contradictions in logical-empirical models. (realism vs. empiricism, analytic vs. synthetic statements, holism, theory-laden observations, scientific explanations, demarcation) (2) Based on this model, the dissertation compares six biological theories from Lamarck (1809), via Cuvier (1811), Geoffroy St. Hilaire (1835), Chambers (1844-60), Owen (1848-68), Wallace (1855/8) to Darwin (1859-1872) and reveals an interesting asymmetry: Compared to any one of his predecessors, Darwins theory appears very original, however, compared to all five predecessor theories, many of these differences disappear and it remains but a small original contribution by Darwin. Thus, Darwin’s is but one in a continuous series of responses to the challenges posed to biology by paleontology and biogeography since the end of the 18th century. (3) A 3-level reception analysis, finally, demonstrates why we speak of a Darwinian revolution nevertheless. (i) A quantitative analysis of nearly 2.000 biological articles reveals that Darwinian concepts where indeed an important theoretical innovation – but definitely not the most important of the time. (ii) When leaving the circle of biology and moving to scientists from other disciplines or educated laymen, the landscape changes. The further outside the biological community, the shallower the audience’s knowledge – and the more visible Darwin’s original contribution. After all, most of Darwin’s contribution can be found in the narrative and worldview of 19th century biology: the only level of knowledge which laymen receive. Evolution Interpretation Rudolf Carnap Ludwig Wittgenstein Beschreibung Historische Epistemologie wissenschaftliche Revolution wissenschaftlicher Fortschritt wissenschaftliche Theorie Kontinuität Diskontinuität wissenschaftliche Erklärungen Abgrenzungskriterium Unterdeterminiertheit analytische Sätze synthetische Sätze Logischer Empirismus wissenschaftlicher Realismus Metaphern Narrative wissenschaftliche Modelle Natürliche Auslese Natürliche Selektion Kampf ums Dasein Homologie Archetype Charles Darwin Robert Chambers Richard Owen Alfred Russel Wallace Georges Cuvier Jean-Baptiste Lamarck Etienne Geoffroy Saint-Hilaire Gottlob Frege Karl Popper evolution interpretation Rudolf Carnap Ludwig Wittgenstein Historical epistemology scientific revolution scientific progress scientific theory continuity discontinuity scientific explanations demarcation criterion underdetermination analytic statements synthetic statements logical empiricism scientific realism metaphors narratives description scientific models natural selection struggle for existence homology archetype Charles Darwin Robert Chambers Richard Owen Alfred Russel Wallace Georges Cuvier Jean-Baptiste Lamarck Etienne Geoffroy Saint-Hilaire Gottlob Frege Karl Popper WH 2000

Page generated in 0.0917 seconds