Author: ARIAS DUART, ANNA
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN ARTIFICIAL INTELLIGENCE
Department: Department of Computer Science (CS)
Mode: Normal
Deposit date: 18/10/2023
Reading date: 11/12/2023
Reading time: 11:00
Reading place: Sala de Juntes de la FIB, Edifici B6, planta 1
Thesis director: CORTÉS GARCÍA, CLAUDIO ULISES | GARCÍA GASULLA, DARIO
Committee:
     PRESIDENT: GIANNOTTI, FOSCA
     SECRETARI: BÉJAR ALONSO, JAVIER
     VOCAL: PADGET, JULIAN
Thesis abstract: Machine learning and, more specifically, deep learning applications have grown in number in recent years. These intelligent systems have shown remarkable performance across various domains, including sensitive areas like medicine and justice. Nevertheless, these models remain opaque, and we need a complete understanding of their internal process. Therefore, the deployment of these black box models can pose risks. Firstly, it might not comply with the current legislation. Secondly, it may lead to severe consequences. Let us consider a scenario in which a model used in a medical application is gender-biased, yielding distinct predictions depending on a person's gender. This fact would perpetuate discrimination against certain parts of the population and exacerbate existing inequalities.To better understand the model's behaviour, enabling the detection and mitigation of potential biases and ultimately achieving more trustworthy models, the eXplainable AI (XAI) field is an active research domain which is growing and receiving increasing attention. Various approaches have been proposed in the literature. Nevertheless, the most widely used are the post-hoc methods. These approaches can be applied once the model is trained, thus preserving the model's original performance. By employing these post-hoc explainability methods to gain insights into the model and identify biases within the datasets and models, we realized that two other biases arise: XAI and human biases.While different XAI methods exist, assessing their faithfulness becomes challenging due to the absence of a ground truth determining what the correct explanation is. The uncertainty regarding whether the explanation accurately reflects the model's behaviour can lead to what we refer to as XAI biases. Is the model biased or is it the explainability method that fails to reflect the model's behaviour? Human bias is another of the biases that emerge when applying these explainability methods. How we show these explanations to humans can be misleading or lead to incorrect conclusions. This can be due to confirmation or automation biases. In addition, when domain experts are asked to review all the explanations, the process can be time-consuming and may lead experts to overlook potential biases in the data and models.The main goal of this thesis is to mitigate the influence of these two new sources of biases (i.e., XAI and human) when explainability is used to detect biases in datasets and models. First, we focus on mitigating XAI biases. To do so, we propose a methodology to assess the reliability of XAI methods. Although our primary goal was to use this methodology within the computer vision discipline, we also demonstrated its applicability in other domains, such as the natural language processing field. After selecting the most reliable XAI method according to our proposed approach, we focus on mitigating human biases. With this objective in mind, we present potential methodologies to semi-automate the detection of data/model biases, thereby reducing the noise introduced by humans. Adopting this approach limits the domain expert's intervention to the final step, in which experts assess whether the biases found are harmful or harmless.

Author: OLEAS SERRANO, DIEGO PATRICIO
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN ARCHITECTURAL DESIGN
Department: Department of Architectural Design (PA)
Mode: Normal
Deposit date: 13/11/2023
Reading date: pending
Reading time: pending
Reading place: pending
Thesis director: FERNANDEZ SALAS, MARIA ELENA | LINARES SOLER, ALFREDO
Committee:
     PRESIDENT: RAVETLLAT MIRA, PEDRO JUAN
     SECRETARI: BARDÍ MILÀ, BERTA
     VOCAL: ALBORNOZ GARCÉS, DIEGO
     VOCAL: PEREIRA FINURAS DE CARVALHO, RICARDO MANUEL
     VOCAL: GRITTI, ANDREA
Thesis abstract: This doctoral thesis examines the process of adaptation of Modern Architecture in Quito over a period of twenty years (1960-1980). The city's location at latitude 0 and an elevation of 2850 meters above sea level determines the unique solar light that influences both the natural and artificial contexts.There are no studies on light under these conditions, nor one that compiles the most important references for those interested in architecture in Quito, especially students. Another motivation is to highlight the efforts of the most influential generation of architects in the country, unknown to most young people, an aspect that this work seeks to address.The methodology includes an extensive bibliographic source, field visits, interviews, and photographic resources, as well as the redrawing of all the projects (38) to establish a common basis for analysis using current technical visualization and modeling resources.The study begins with a description of the Andean mountains in general, the Andes in Ecuador, and concludes with the Guayllabamba Basin, where the city of Quito is located, along with its volcano, Pichincha. This first part is complemented by the research of Alexander von Humboldt, who in the 19th century led an expedition to study the flora and fauna of various Latin American countries, including Ecuador.The second part of the thesis presents the city of Quito in three moments: the pre-Columbian city, the colonial city, and the modern city.In the third part, the influence of Humboldt on landscape painting is revisited, followed by landscaping in Ecuador and painting in the mid-20th century. Painting, along with photography and watercolor, contributes to understanding the light of Quito, as well as the particular descriptions of it in literature.The following section organizes specific aspects of light in terms of principles, spaces, and relevant lighting elements in the urban environment of Quito.In the fourth section, the process of learning and adaptation of modern architecture in Quito is addressed.The analysis of three selected works for their relationship with the landscape reveals the value of the study. As a complement and context to these works, the systematic analysis of 38 works considered in this work is presented, serving as a very useful reference.In conclusion, this doctoral thesis highlights the importance of a comprehensive approach in architecture, focused on light and the environment. The findings of this work provide valuable information for architecture in harmony with Quito's environment, promoting identity and suggesting possible directions for future research.

Author: OCKOVA, JANA
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN PHOTONICS
Department: Institute of Photonic Sciences (ICFO)
Mode: Normal
Deposit date: 06/11/2023
Reading date: 13/12/2023
Reading time: 10:00
Reading place: ICFO, Mediterranean Technology Park, Avinguda Carl Friedrich Gauss, 3, 08860 Castelldefels, Barcelona
Thesis director: VAN HULST, NIEK
Committee:
     PRESIDENT: GARCÍA PARAJO, MARÍA
     SECRETARI: WIENTJES, INA EMILIA
     VOCAL: POTMA, ERIC OLAF
Thesis abstract: Light is a powerful non invasive tool for probing matter down to its fundamental molecular properties. The past three decades saw advent of metallic nanoantennas engineered to concentrate light into sub diffraction limited hotspots and enhance optical properties of nearby emitters by many orders of magnitude. This boosted optical microscopy, allowing it to interrogate even extremely dim systems at their most fundamental single molecule level. The enhancement and local confinement also unlocked sensing applications down to zeptomolar concentrations, which can revolutionise environmental monitoring, clinical diagnosis and personalised medicine. Beyond sensing, metallic nanoparticles can improve the efficiency of photovoltaic devices and next generation green catalysts. The current challenge for large scale practical implementations is lack of understanding and control of the underlying nanoscale processes. Here, we use optical microscopy and metallic nanoantennas to perform single molecule and single particle experiments to shed light on fundamental mechanism of photosynthesis, nanoscale parameters crucial for sensing and underlying photochemistry in nanoantenna hotspots relevant for catalysis.Firstly, we employ gold nanorods and cryomicroscopy to study excitation energy transfer in the Fenna Matthews Olson photosynthetic complex. By probing one complex at a time at room temperature and 77 K, we uncover energy transfer between its subunits, where both experimental approaches constitute the first of their kind for this extremely dim system. Furthermore, we show that maximising the nanorod enhancement likely yields more efficient energy transfer to the nanorod than between the subunits of the complex, making them operate as effectively independent. Our results shed new light on the role of excitation transfer and annihilation in the regulation of photosynthesis.Next, we evaluate Raman scattering enhancement of a library of ten nanoparticles using a home built automated Raman microscope. By recording a statistically significant dataset of spectral traces from discrete nanoscale spots, we can distinguish Raman enhancement performance of different types of nanoparticles that would otherwise appear identical in classical bulk measurements. Furthermore, adding a dark field scattering detection allows us to classify the measurements between single and multiple nanoparticles and directly probe the variability of single particle enhancements. This is a crucial parameter for sensing applications and the detailed nanoscale insight provided by our measurement platform can be used to accelerate the rational design of new nanoparticles for quantitative sensing.Finally, we employ the automated Raman microscope to study light induced chemical reactions in metallic nanocavities. Specifically, we record surface enhanced Raman scattering of a few methylene blue molecules sandwiched between a gold mirror and a gold nanoparticle. We develop a new sample assembly compatible with oil immersion that yields a 150 fold increase in the molecular signal than previously published air coupling schemes. We use a pulsed laser to induce a chemical transformation of the methylene blue molecules. By interpreting the results in the context of plasmonic properties of the gold nanojunction obtained from dark field measurements and simulations, we were able to rule out lattice heating and narrow down the underlying mechanism to a plasmon induced sub picosecond process. Furthermore, we propose that spontaneous picosecond Raman spectroscopy is suitable to study reactions at metallic surfaces which lie at the heart of heterogeneous catalysis.

Author: RAMÍREZ SANTOS, ANA GLADYS
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN SUSTAINABILITY
Department: University Research Institute for Sustainability Science and Technology (IS.UPC)
Mode: Normal
Deposit date: 02/10/2023
Reading date: 13/12/2023
Reading time: 09:00
Reading place: Sala Conferencias TR1, ESEIAAT, Campus de Terrassa
Thesis director: MORATO FARRERAS, JORDI | RIVERA FERRE, MARTA GUADALUPE
Committee:
     PRESIDENT: CALVET-MIR, LAURA
     SECRETARI: ALVAREZ DEL CASTILLO, JAVIER
     VOCAL NO PRESENCIAL: LÓPEZ-GARCÍA, DANIEL
Thesis abstract: This doctoral research develops the issue of just agroecological transitions in the food system and develops two case studies in the Global North and Global South. Based on the framework developed to analyze sustainable transitions, i.e., the theoretical framework of the multilevel perspective (MLP) is considered one of the most widely used frameworks of analysis to understand the multidimensional complexity of transitions. Transitions in MLP are defined as the change from one Regime to another, and as the result of the interaction of three levels of action: Niche-Regime-Landscape. Based on this approach, works have been developed on the analysis of how the change promoted by transitions from one system to another can generate profound social change and at the same time new social and environmental injustices and vulnerabilities. Also, based on the MLP the groecological transitions approach is developed, offering an integration of knowledge, combining science, movement, and practice to achieve sustainable, equitable and just food systems. In both theoretical proposals, elements for analysis converge in relation to traditional agroecological knowledge (epistemic justice) and gender approaches (gender justice). In the case of this doctoral dissertation research, we consider the epistemic justice and gender justice two elements indispensable and fundamental to integrate in the analysis of the processes aimed at transforming food systems towards more sustainable, healthy, and just ones. In addition, in order to understand and identify how the dynamics between gender and traditional agroecological knowledge have been explored, the critical elements that influence the adaptation, modification and transmission of knowledge, and to make visible the different experiences that men and women have within agri-food systems through a gender lens, a literature review is developed presenting diverse case studies at the Global level. Finally, based on the theoretical perspective of MLP, just transitions and agroecological transitions, two case studies are addressed. The first one located in Santa Isabel Cholula, Puebla, Mexico, Global South, where the main challenges and enablers of a just agroecological transition within the agrofood systems are explored. Also, we identify the traditional knowledge and its characteristics according to gender. The second case study is located in the city of Barcelona, Global North, where it is explored how the Niches-Regime-Landscape of the city interact for the transformation of the food system, as well as the way in which the gender perspective, knowledge and feminism are addressed within the actions between the Niches and the Regime. Key words: Food, system, Sustainable agrifood systems, sustainable transition, agroecological transition, just transition, traditional agroecological knowledge, epistemic justice, gender justice, feminism.

Author: SIDKI RIUS, NOR
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN NATURAL RESOURCES AND THE ENVIRONMENT
Department: Department of Mining Engineering and Natural Resources (EMRN)
Mode: Normal
Deposit date: 02/11/2023
Reading date: 13/12/2023
Reading time: 11:30
Reading place: Sala d'Actes EPSEM (Av.Bases de Manresa, 61-73, 08242 Manresa)
Thesis director: SANMIQUEL PERA, LLUIS | BASCOMPTA MASSANÈS, MARC
Committee:
     PRESIDENT: RODRÍGUEZ DÍEZ, RAFAEL
     SECRETARI: OLIVA MONCUNILL, JOSEP
     VOCAL: CASTAÑÓN GARCÍA, ANA
Thesis abstract: The importance of potash is inevitable to life on the planet earth as it is required by living mechanisms including plants,animals and human beings. In Europe, this vital ore is produced mainly by Germany and Spain, which consists of up to 10%of the global supply. As a matter of fact, higher demand for potash may be required due to population growth. This may leadto increasing potash mining activities, which might create instability with regard to the geomechanical features of the miningareas and potential environmental and social impacts.The main objective of this research is to approach innovative models that can describe the geomechanical behavior of theCatalan Potash Basin deposit. Including concern about surface subsidence, high deformations and collapses ofunderground potash mining, with the ultimate goal of improving the present working conditions and mitigating theenvironmental impacts. Four Specific objectives have been defined; (1) The creation of a model to predict subsidence basedon the geomechanical parameters of the ore deposit. (2) The Geochemical characterization of the pelitic and insolublephase detected between the saline beds and within them. (3) The characterization of the surface subsidence basin,including boundary angle and distance of influence. (4) The assessment of the deformations in the underground miningfacilities. To achieve these objectives, a combination of FEM, InSAR, GPS, Laser Scanner and Geochemical techniques hasbeen used to predict the subsidence and achieve a better geomechanical behavior characterization of it, which have resultedin two main outcomes. First, it has been established that numerical 2D modeling is a useful tool for planning landmanagement and mining operations, providing the flexibility and predictability needed to manage huge surfaces and reduceenvironmental consequences. The approach is also adaptable to other subsurface infrastructure types. Second, thesubsidence basin has been successfully characterized. Third, pelitic layers with insoluble mineralogy within salinelithologies have successfully undergone geochemical examination. However, to expand the study area to include the entireactive ore deposit mineralogy, additional examination will be required. Fourth, the laser scanner technique has beensatisfactorily assessed, it has given important insights into the displacements that occur during the early opening of anunderground mining drift affirming that it is a suitable technique to be implemented for routinely tracking deformations in thegeomechanical management of the ore deposit. However, it is necessary to perform a precise topographic control of thetargets used, in order to correctly georeference the point clouds produced by the laser scanner. Furthermore, characterizationof the subsidence has been successfully assessed in the eastern portion of the Catalan Potash Basin. With the help of themethods mentioned, it was possible to evaluate a large dataset and identify four unique zones, one of which wasexperiencing residual subsidence. Mining firms may choose to use this flexible and effective method and apply it to otherstudies involving subsidence. In the end, based on the results and objectives achieved in the study, suggestions for furtherresearch have been proposed, which can further advance the knowledge in this line of research.The structure of the thesis is as follows: Chapter 1 introduces the case study and states the objectives and hypothesisdeveloped in the research. In chapter two, an extensive and precise bibliographic review is made. Chapter 3 describes themethodologies followed in the different investigations developed. Chapter 4 presents the results and discusses them.Finally, Chapter 5 presents the conclusions and proposes future research to improve and complement the knowledgeobtained