Mitigating non-Lambertian surfaces issues in Stereo Matching with Neural Radiance Fields

Depth estimation from images has long been regarded as a preferable alternative compared to expensive and intrusive active sensors, such as LiDAR and ToF. The topic has attracted the attention of an increasingly wide audience thanks to the great amount of application domains, such as autonomous driving, robotic navigation and 3D reconstruction. Among the various techniques employed for depth estimation, stereo matching is one of the most widespread, owing to its robustness, speed and simplicity in setup. Recent developments has been aided by the abundance of annotated stereo images, which granted to deep learning the opportunity to thrive in a research area where deep networks can reach state-of-the-art sub-pixel precision in most cases. Despite the recent findings, stereo matching still begets many open challenges, two among them being finding pixel correspondences in presence of objects that exhibits a non-Lambertian behaviour and processing high-resolution images. Recently, a novel dataset named Booster, which contains high-resolution stereo pairs featuring a large collection of labeled non-Lambertian objects, has been released. The work shown that training state-of-the-art deep neural network on such data improves the generalization capabilities of these networks also in presence of non-Lambertian surfaces. Regardless being a further step to tackle the aforementioned challenge, Booster includes a rather small number of annotated images, and thus cannot satisfy the intensive training requirements of deep learning. This thesis work aims to investigate novel view synthesis techniques to augment the Booster dataset, with ultimate goal of improving stereo matching reliability in presence of high-resolution images that displays non-Lambertian surfaces.

Neural Radiance Fields for Relighting and View Synthesis

Neural scene representation and neural rendering are new computer vision techniques that enable the reconstruction and implicit representation of real 3D scenes from a set of 2D captured images, by fitting a deep neural network. The trained network can then be used to render novel views of the scene. A recent work in this field, Neural Radiance Fields (NeRF), presented a state-of-the-art approach, which uses a simple Multilayer Perceptron (MLP) to generate photo-realistic RGB images of a scene from arbitrary viewpoints. However, NeRF does not model any light interaction with the fitted scene; therefore, despite producing compelling results for the view synthesis task, it does not provide a solution for relighting. In this work, we propose a new architecture to enable relighting capabilities in NeRF-based representations and we introduce a new real-world dataset to train and evaluate such a model. Our method demonstrates the ability to perform realistic rendering of novel views under arbitrary lighting conditions.

Modeling, characterization and microalgal cultivation in an airlift panel photobioreactors.

Microalgae cultures are attracting great attentions in many industrial applications. However, one of the technical challenges is to cut down the capital and operational costs of microalgae production systems, with special difficulty in reactor design and scale-up. The thesis work open with an overview on the microalgae cultures as a possible answer to solve some of the upcoming planet issues and their applications in several fields. After the work offers a general outline on the state of the art of microalgae culture systems, taking a special look to the enclosed photobioreactors (PBRs). The overall objective of this study is to advance the knowledge of PBRs design and lead to innovative large scale processes of microalgae cultivation. An airlift flat panel photobioreactor was designed, modeled and experimentally characterized. The gas holdup, liquid flow velocity and oxygen mass transfer of the reactor were experimentally determined and mathematically modeled, and the performance of the reactor was tested by cultivation of microalgae. The model predicted data correlated well with experimental data, and the high concentration of suspension cell culture could be achieved with controlled conditions. The reactor was inoculated with the algal strain Scenedesmus obliquus sp. first and with Chlorella sp. later and sparged with air. The reactor was operated in batch mode and daily monitored for pH, temperature, and biomass concentration and activity. The productivity of the novel device was determined, suggesting the proposed design can be effectively and economically used in carbon dioxide mitigation technologies and in the production of algal biomass for biofuel and other bioproducts. Those research results favored the possibility of scaling the reactor up into industrial scales based on the models employed, and the potential advantages and disadvantages were discussed for this novel industrial design.

Finite element modeling of flow/compression-induced deformation of alginate scaffolds for bone tissue engineering

Trauma or degenerative diseases such as osteonecrosis may determine bone loss whose recover is promised by a "tissue engineering“ approach. This strategy involves the use of stem cells, grown onboard of adequate biocompatible/bioreabsorbable hosting templates (usually defined as scaffolds) and cultured in specific dynamic environments afforded by differentiation-inducing actuators (usually defined as bioreactors) to produce implantable tissue constructs. The purpose of this thesis is to evaluate, by finite element modeling of flow/compression-induced deformation, alginate scaffolds intended for bone tissue engineering. This work was conducted at the Biomechanics Laboratory of the Institute of Biomedical and Neural Engineering of the Reykjavik University of Iceland. In this respect, Comsol Multiphysics 5.1 simulations were carried out to approximate the loads over alginate 3D matrices under perfusion, compression and perfusion+compression, when varyingalginate pore size and flow/compression regimen. The results of the simulations show that the shear forces in the matrix of the scaffold increase coherently with the increase in flow and load, and decrease with the increase of the pore size. Flow and load rates suggested for proper osteogenic cell differentiation are reported.

A deep learning model via thermodynamics and condensed matter physics

Deep Learning architectures give brilliant results in a large variety of fields, but a comprehensive theoretical description of their inner functioning is still lacking. In this work, we try to understand the behavior of neural networks by modelling in the frameworks of Thermodynamics and Condensed Matter Physics. We approach neural networks as in a real laboratory and we measure the frequency spectrum and the entropy of the weights of the trained model. The stochasticity of the training occupies a central role in the dynamics of the weights and makes it difficult to assimilate neural networks to simple physical systems. However, the analogy with Thermodynamics and the introduction of a well defined temperature leads us to an interesting result: if we eliminate from a CNN the "hottest" filters, the performance of the model remains the same, whereas, if we eliminate the "coldest" ones, the performance gets drastically worst. This result could be exploited in the realization of a training loop which eliminates the filters that do not contribute to loss reduction. In this way, the computational cost of the training will be lightened and more importantly this would be done by following a physical model. In any case, beside important practical applications, our analysis proves that a new and improved modeling of Deep Learning systems can pave the way to new and more efficient algorithms.

Neural Network Based UAV Fault Tolerant Control with Aerodynamic Disturbance Estimation and Compensation

In this thesis, the problem of controlling a quadrotor UAV is considered. It is done by presenting an original control system, designed as a combination of Neural Networks and Disturbance Observer, using a composite learning approach for a system of the second order, which is a novel methodology in literature. After a brief introduction about the quadrotors, the concepts needed to understand the controller are presented, such as the main notions of advanced control, the basic structure and design of a Neural Network, the modeling of a quadrotor and its dynamics. The full simulator, developed on the MATLAB Simulink environment, used throughout the whole thesis, is also shown. For the guidance and control purposes, a Sliding Mode Controller, used as a reference, it is firstly introduced, and its theory and implementation on the simulator are illustrated. Finally the original controller is introduced, through its novel formulation, and implementation on the model. The effectiveness and robustness of the two controllers are then proven by extensive simulations in all different conditions of external disturbance and faults.

Investigating perceptual multisensory impairments in Autism through a neural network: a possible neural implementation for the shift from cross-modal competition to facilitation

Resolution of multisensory deficits has been observed in teenagers with Autism Spectrum Disorders (ASD) for complex, social speech stimuli; this resolution extends to more basic multisensory processing, involving low-level stimuli. In particular, a delayed transition of multisensory integration (MSI) from a default state of competition to one of facilitation has been observed in ASD children. In other terms, the complete maturation of MSI is achieved later in ASD. In the present study a neuro-computational model is used to reproduce some patterns of behavior observed experimentally, modeling a bisensory reaction time task, in which auditory and visual stimuli are presented in random sequence alone (A or V) or together (AV). The model explains how the default competitive state can be implemented via mutual inhibition between primary sensory areas, and how the shift toward the classical multisensory facilitation, observed in adults, is the result of inhibitory cross-modal connections becoming excitatory during the development. Model results are consistent with a stronger cross-modal inhibition in ASD children, compared to normotypical (NT) ones, suggesting that the transition toward a cooperative interaction between sensory modalities takes longer to occur. Interestingly, the model also predicts the difference between unisensory switch trials (in which sensory modality switches) and unisensory repeat trials (in which sensory modality repeats). This is due to an inhibitory mechanism, characterized by a slow dynamics, driven by the preceding stimulus and inhibiting the processing of the incoming one, when of the opposite sensory modality. These findings link the cognitive framework delineated by the empirical results to a plausible neural implementation.

Bio-QA-GNN: Reasoning with Language Models and Knowledge Graphs for Interpretable Biomedical Question Answering

Nowadays the idea of injecting world or domain-specific structured knowledge into pre-trained language models (PLMs) is becoming an increasingly popular approach for solving problems such as biases, hallucinations, huge architectural sizes, and explainability lack—critical for real-world natural language processing applications in sensitive fields like bioinformatics. One recent work that has garnered much attention in Neuro-symbolic AI is QA-GNN, an end-to-end model for multiple-choice open-domain question answering (MCOQA) tasks via interpretable text-graph reasoning. Unlike previous publications, QA-GNN mutually informs PLMs and graph neural networks (GNNs) on top of relevant facts retrieved from knowledge graphs (KGs). However, taking a more holistic view, existing PLM+KG contributions mainly consider commonsense benchmarks and ignore or shallowly analyze performances on biomedical datasets. This thesis start from a propose of a deep investigation of QA-GNN for biomedicine, comparing existing or brand-new PLMs, KGs, edge-aware GNNs, preprocessing techniques, and initialization strategies. By combining the insights emerged in DISI's research, we introduce Bio-QA-GNN that include a KG. Working with this part has led to an improvement in state-of-the-art of MCOQA model on biomedical/clinical text, largely outperforming the original one (+3.63\% accuracy on MedQA). Our findings also contribute to a better understanding of the explanation degree allowed by joint text-graph reasoning architectures and their effectiveness on different medical subjects and reasoning types. Codes, models, datasets, and demos to reproduce the results are freely available at: \url{https://github.com/disi-unibo-nlp/bio-qagnn}.

Numerical modeling on the short-term response of fiber-reinforced concrete round panels

Fiber-reinforced concrete is a composite material consisting of discrete, discontinuous, and uniformly distributed fibers in plain concrete primarily used to enhance the tensile properties of the concrete. FRC performance depends upon the fiber, interface, and matrix properties. The use of fiber-reinforced concrete has been increasing substantially in the past few years in different fields of the construction industry such as ground-level application in sidewalks and building floors, tunnel lining, aircraft parking, runways, slope stabilization, etc. Many experiments have been performed to observe the short-term and long-term mechanical behavior of fiber-reinforced concrete in the last decade and numerous numerical models have been formulated to accurately capture the response of fiber-reinforced concrete. The main purpose of this dissertation is to numerically calibrate the short-term response of the concrete and fiber parameters in mesoscale for the three-point bending test and cube compression test in the MARS framework which is based on the lattice discrete particle model (LDPM) and later validate the same parameters for the round panels. LDPM is the most validated theory in mesoscale theories for concrete. Different seeds representing the different orientations of concrete and fiber particles are simulated to produce the mean numerical response. The result of numerical simulation shows that the lattice discrete particle model for fiber-reinforced concrete can capture results of experimental tests on the behavior of fiber-reinforced concrete to a great extent.