Development and characterization of deep learning techniques for neuroimaging data

Deep learning methods are extremely promising machine learning tools to analyze neuroimaging data. However, their potential use in clinical settings is limited because of the existing challenges of applying these methods to neuroimaging data. In this study, first a data leakage type caused by slice-level data split that is introduced during training and validation of a 2D CNN is surveyed and a quantitative assessment of the model’s performance overestimation is presented. Second, an interpretable, leakage-fee deep learning software written in a python language with a wide range of options has been developed to conduct both classification and regression analysis. The software was applied to the study of mild cognitive impairment (MCI) in patients with small vessel disease (SVD) using multi-parametric MRI data where the cognitive performance of 58 patients measured by five neuropsychological tests is predicted using a multi-input CNN model taking brain image and demographic data. Each of the cognitive test scores was predicted using different MRI-derived features. As MCI due to SVD has been hypothesized to be the effect of white matter damage, DTI-derived features MD and FA produced the best prediction outcome of the TMT-A score which is consistent with the existing literature. In a second study, an interpretable deep learning system aimed at 1) classifying Alzheimer disease and healthy subjects 2) examining the neural correlates of the disease that causes a cognitive decline in AD patients using CNN visualization tools and 3) highlighting the potential of interpretability techniques to capture a biased deep learning model is developed. Structural magnetic resonance imaging (MRI) data of 200 subjects was used by the proposed CNN model which was trained using a transfer learning-based approach producing a balanced accuracy of 71.6%. Brain regions in the frontal and parietal lobe showing the cerebral cortex atrophy were highlighted by the visualization tools.

Multi-Principal Agency and the Compliance with Professional Guidelines in Health Systems: Economic and Regulatory Perspectives

Amid the trend of rising health expenditure in developed economies, changing the healthcare delivery models is an important point of action for service regulators to contain this trend. Such a change is mostly induced by either financial incentives or regulatory tools issued by the regulators and targeting service providers and patients. This creates a tripartite interaction between service regulators, professionals, and patients that manifests a multi-principal agent relationship, in which professionals are agents to two principals: regulators and patients. This thesis is concerned with such a multi-principal agent relationship in healthcare and attempts to investigate the determinants of the (non-)compliance to regulatory tools in light of this tripartite relationship. In addition, the thesis provides insights into the different institutional, economic, and regulatory settings, which govern the multi-principal agent relationship in healthcare in different countries. Furthermore, the thesis provides and empirically tests a conceptual framework of the possible determinants of (non-)compliance by physicians to regulatory tools issued by the regulator. The main findings of the thesis are first, in a multi-principal agent setting, the utilization of financial incentives to align the objectives of professionals and the regulator is important but not the only solution. This finding is based on the heterogeneity in the financial incentives provided to professionals in different health markets, which does not provide a one-size-fits-all model of financial incentives to influence clinical decisions. Second, soft law tools as clinical practice guidelines (CPGs) are important tools to mitigate the problems of the multi-principal agent setting in health markets as they reduce information asymmetries while preserving the autonomy of professionals. Third, CPGs are complex and heterogeneous and so are the determinants of (non-)compliance to them. Fourth, CPGs work but under conditions. Factors such as intra-professional competition between service providers or practitioners might lead to non-compliance to CPGs – if CPGs are likely to reduce the professional’s utility. Finally, different degrees of soft law mandate have different effects on providers’ compliance. Generally, the stronger the mandate, the stronger the compliance, however, even with a strong mandate, drivers such as intra-professional competition and co-management of patients by different professionals affected the (non-)compliance.

Unsupervised reinforcement learning via state entropy maximization

Reinforcement Learning (RL) provides a powerful framework to address sequential decision-making problems in which the transition dynamics is unknown or too complex to be represented. The RL approach is based on speculating what is the best decision to make given sample estimates obtained from previous interactions, a recipe that led to several breakthroughs in various domains, ranging from game playing to robotics. Despite their success, current RL methods hardly generalize from one task to another, and achieving the kind of generalization obtained through unsupervised pre-training in non-sequential problems seems unthinkable. Unsupervised RL has recently emerged as a way to improve generalization of RL methods. Just as its non-sequential counterpart, the unsupervised RL framework comprises two phases: An unsupervised pre-training phase, in which the agent interacts with the environment without external feedback, and a supervised fine-tuning phase, in which the agent aims to efficiently solve a task in the same environment by exploiting the knowledge acquired during pre-training. In this thesis, we study unsupervised RL via state entropy maximization, in which the agent makes use of the unsupervised interactions to pre-train a policy that maximizes the entropy of its induced state distribution. First, we provide a theoretical characterization of the learning problem by considering a convex RL formulation that subsumes state entropy maximization. Our analysis shows that maximizing the state entropy in finite trials is inherently harder than RL. Then, we study the state entropy maximization problem from an optimization perspective. Especially, we show that the primal formulation of the corresponding optimization problem can be (approximately) addressed through tractable linear programs. Finally, we provide the first practical methodologies for state entropy maximization in complex domains, both when the pre-training takes place in a single environment as well as multiple environments.

Real-time anomaly detection of gamma-ray bursts for the Cherenkov Telescope Array using deep learning

The Cherenkov Telescope Array (CTA) will be the next-generation ground-based observatory to study the universe in the very-high-energy domain. The observatory will rely on a Science Alert Generation (SAG) system to analyze the real-time data from the telescopes and generate science alerts. The SAG system will play a crucial role in the search and follow-up of transients from external alerts, enabling multi-wavelength and multi-messenger collaborations. It will maximize the potential for the detection of the rarest phenomena, such as gamma-ray bursts (GRBs), which are the science case for this study. This study presents an anomaly detection method based on deep learning for detecting gamma-ray burst events in real-time. The performance of the proposed method is evaluated and compared against the Li&Ma standard technique in two use cases of serendipitous discoveries and follow-up observations, using short exposure times. The method shows promising results in detecting GRBs and is flexible enough to allow real-time search for transient events on multiple time scales. The method does not assume background nor source models and doe not require a minimum number of photon counts to perform analysis, making it well-suited for real-time analysis. Future improvements involve further tests, relaxing some of the assumptions made in this study as well as post-trials correction of the detection significance. Moreover, the ability to detect other transient classes in different scenarios must be investigated for completeness. The system can be integrated within the SAG system of CTA and deployed on the onsite computing clusters. This would provide valuable insights into the method's performance in a real-world setting and be another valuable tool for discovering new transient events in real-time. Overall, this study makes a significant contribution to the field of astrophysics by demonstrating the effectiveness of deep learning-based anomaly detection techniques for real-time source detection in gamma-ray astronomy.

Intermediate linguistic task fine-tuning on Multi-lingual models

State-of-the-art NLP systems are generally based on the assumption that the underlying models are provided with vast datasets to train on. However, especially when working in multi-lingual contexts, datasets are often scarce, thus more research should be carried out in this field. This thesis investigates the benefits of introducing an additional training step when fine-tuning NLP models, named Intermediate Training, which could be exploited to augment the data used for the training phase. The Intermediate Training step is applied by training models on NLP tasks that are not strictly related to the target task, aiming to verify if the models are able to leverage the learned knowledge of such tasks. Furthermore, in order to better analyze the synergies between different categories of NLP tasks, experimentations have been extended also to Multi-Task Training, in which the model is trained on multiple tasks at the same time.

Graph-based Vision-Language Model for Multi-Modal Retrieval of Fashion Products

Artificial Intelligence is reshaping the field of fashion industry in different ways. E-commerce retailers exploit their data through AI to enhance their search engines, make outfit suggestions and forecast the success of a specific fashion product. However, it is a challenging endeavour as the data they possess is huge, complex and multi-modal. The most common way to search for fashion products online is by matching keywords with phrases in the product's description which are often cluttered, inadequate and differ across collections and sellers. A customer may also browse an online store's taxonomy, although this is time-consuming and doesn't guarantee relevant items. With the advent of Deep Learning architectures, particularly Vision-Language models, ad-hoc solutions have been proposed to model both the product image and description to solve this problems. However, the suggested solutions do not exploit effectively the semantic or syntactic information of these modalities, and the unique qualities and relations of clothing items. In this work of thesis, a novel approach is proposed to address this issues, which aims to model and process images and text descriptions as graphs in order to exploit the relations inside and between each modality and employs specific techniques to extract syntactic and semantic information. The results obtained show promising performances on different tasks when compared to the present state-of-the-art deep learning architectures.

Hierarchical Multi-Label Text Classification in a Low-Resource Setting

In this thesis we address a multi-label hierarchical text classification problem in a low-resource setting and explore different approaches to identify the best one for our case. The goal is to train a model that classifies English school exercises according to a hierarchical taxonomy with few labeled data. The experiments made in this work employ different machine learning models and text representation techniques: CatBoost with tf-idf features, classifiers based on pre-trained models (mBERT, LASER), and SetFit, a framework for few-shot text classification. SetFit proved to be the most promising approach, achieving better performance when during training only a few labeled examples per class are available. However, this thesis does not consider all the hierarchical taxonomy, but only the first two levels: to address classification with the classes at the third level further experiments should be carried out, exploring methods for zero-shot text classification, data augmentation, and strategies to exploit the hierarchical structure of the taxonomy during training.

Graphene And Carbon Nanotube Nanocomposite For Gene Transfection.

Graphene and carbon nanotube nanocomposite (GCN) was synthesised and applied in gene transfection of pIRES plasmid conjugated with green fluorescent protein (GFP) in NIH-3T3 and NG97 cell lines. The tips of the multi-walled carbon nanotubes (MWCNTs) were exfoliated by oxygen plasma etching, which is also known to attach oxygen content groups on the MWCNT surfaces, changing their hydrophobicity. The nanocomposite was characterised by high resolution scanning electron microscopy; energy-dispersive X-ray, Fourier transform infrared and Raman spectroscopies, as well as zeta potential and particle size analyses using dynamic light scattering. BET adsorption isotherms showed the GCN to have an effective surface area of 38.5m(2)/g. The GCN and pIRES plasmid conjugated with the GFP gene, forming π-stacking when dispersed in water by magnetic stirring, resulting in a helical wrap. The measured zeta potential confirmed that the plasmid was connected to the nanocomposite. The NIH-3T3 and NG97 cell lines could phagocytize this wrap. The gene transfection was characterised by fluorescent protein produced in the cells and pictured by fluorescent microscopy. Before application, we studied GCN cell viability in NIH-3T3 and NG97 line cells using both MTT and Neutral Red uptake assays. Our results suggest that GCN has moderate stability behaviour as colloid solution and has great potential as a gene carrier agent in non-viral based therapy, with low cytotoxicity and good transfection efficiency.

Doehlert Design-desirability Function Multi-criteria Optimal Separation Of 17 Phenolic Compounds From Extra-virgin Olive Oil By Capillary Zone Electrophoresis.

In Brazil, the consumption of extra-virgin olive oil (EVOO) is increasing annually, but there are no experimental studies concerning the phenolic compound contents of commercial EVOO. The aim of this work was to optimise the separation of 17 phenolic compounds already detected in EVOO. A Doehlert matrix experimental design was used, evaluating the effects of pH and electrolyte concentration. Resolution, runtime and migration time relative standard deviation values were evaluated. Derringer's desirability function was used to simultaneously optimise all 37 responses. The 17 peaks were separated in 19min using a fused-silica capillary (50μm internal diameter, 72cm of effective length) with an extended light path and 101.3mmolL(-1) of boric acid electrolyte (pH 9.15, 30kV). The method was validated and applied to 15 EVOO samples found in Brazilian supermarkets.

Ecological Literacy And Beyond: Problem-based Learning For Future Professionals.

Ecological science contributes to solving a broad range of environmental problems. However, lack of ecological literacy in practice often limits application of this knowledge. In this paper, we highlight a critical but often overlooked demand on ecological literacy: to enable professionals of various careers to apply scientific knowledge when faced with environmental problems. Current university courses on ecology often fail to persuade students that ecological science provides important tools for environmental problem solving. We propose problem-based learning to improve the understanding of ecological science and its usefulness for real-world environmental issues that professionals in careers as diverse as engineering, public health, architecture, social sciences, or management will address. Courses should set clear learning objectives for cognitive skills they expect students to acquire. Thus, professionals in different fields will be enabled to improve environmental decision-making processes and to participate effectively in multidisciplinary work groups charged with tackling environmental issues.

Behavior Of Listeria Monocytogenes In A Multi-species Biofilm With Enterococcus Faecalis And Enterococcus Faecium And Control Through Sanitation Procedures.

The formation of mono-species biofilm (Listeria monocytogenes) and multi-species biofilms (Enterococcus faecium, Enterococcus faecalis, and L. monocytogenes) was evaluated. In addition, the effectiveness of sanitation procedures for the control of the multi-species biofilm also was evaluated. The biofilms were grown on stainless steel coupons at various incubation temperatures (7, 25 and 39°C) and contact times (0, 1, 2, 4, 6 and 8days). In all tests, at 7°C, the microbial counts were below 0.4 log CFU/cm(2) and not characteristic of biofilms. In mono-species biofilm, the counts of L. monocytogenes after 8days of contact were 4.1 and 2.8 log CFU/cm(2) at 25 and 39°C, respectively. In the multi-species biofilms, Enterococcus spp. were present at counts of 8 log CFU/cm(2) at 25 and 39°C after 8days of contact. However, the L. monocytogenes in multi-species biofilms was significantly affected by the presence of Enterococcus spp. and by temperature. At 25°C, the growth of L. monocytogenes biofilms was favored in multi-species cultures, with counts above 6 log CFU/cm(2) after 8days of contact. In contrast, at 39°C, a negative effect was observed for L. monocytogenes biofilm growth in mixed cultures, with a significant reduction in counts over time and values below 0.4 log CFU/cm(2) starting at day 4. Anionic tensioactive cleaning complemented with another procedure (acid cleaning, disinfection or acid cleaning+disinfection) eliminated the multi-species biofilms under all conditions tested (counts of all micro-organisms<0.4 log CFU/cm(2)). Peracetic acid was the most effective disinfectant, eliminating the multi-species biofilms under all tested conditions (counts of the all microorganisms <0.4 log CFU/cm(2)). In contrast, biguanide was the least effective disinfectant, failing to eliminate biofilms under all the test conditions.

Autoscopia: um procedimento de pesquisa e de formação

This article considers a procedure for data collection called autoscopy. Autoscopy entails the video recording of a practice with the purpose of allowing analysis and self-evaluation by one of the protagonists of that practice. The objective of the video recording is that of apprehending the actions of the agent (or agents), the scenario, and the plot that make up a situation. The recorded material is subjected to sessions of analysis after the action that aim at the understanding of the reflective process of the agent (or agents) through their verbalizations during the analysis of video recorded scenes. The present text introduces a theoretical basis for the procedure of autoscopy, deals with advantages and limitations of its use, as well as with aspects that deserve attention and, finally, describes the authors' experiences in two studies in which the procedure was employed. Starting from these two experiences, differences and similarities are pointed out between the studies, especially regarding the participants, object, and the time distribution of the video recordings. The authors draw considerations about the formative-reflective potential of the procedure, both for research situations and for the learning and training of various professionals, considering it to be an excellent educational instrument. It is, however, vital to keep in mind the need to recognize and return to the teacher, as an autoscopic participant, his condition as subject of his own profession, thereby promoting, besides the self-evaluation, also the autonomy of his thinking and doing.

A língua inglesa no ensino fundamentl I público: diálogos com Bakhtin por uma formação plurilíngue

This article aims at discussing the contributions of the Bakhtinian Circle theories to foreign language teaching and learning (HALL et al., 2005), as far as the first years of formal education in Brazil are concerned. Up to the present moment, foreign languages, including English, are not officially part of the National Curriculum of the first five schooling years. Due to the importance of English in a globalized world and despite all the controversial socio-educational impacts of such an influence, there has been an increase in the interest in this discipline at the beginning years of Brazilian public education (ROCHA, 2006), which has been happening at an irregular pace and without official parameters. Therefore, the relevance of this work lies on the possible guidelines it may offer to support a more effective, situated and meaningful teaching-learning process in that context. Standing for a pluralistic approach to language education, we take the bakhtinian speech genres as organizers of the educational process. We strongly believe that through a dialogic, pluralistic and trans/intercultural teaching (MAHER, 2007), whose main objective is the development of multi (COPE e KALANTZIS, 2000) and critical (COMBER, 2006) literacies, the hybridization of genres and cultures, as well as the creation of third spaces (KOSTOGRIZ, 2005; KUMARAVADIVELU, 2008) can happen. From this perspective, foreign language teaching and learning play a transformative role in society and English is seen as a boundary object (STAR e GRIESEMER, 1989), in and by which diversity, pluralism and polyphony can naturally find their way.

Flicker perimetry in healthy subjects: influence of age and gender, learning effect and short-term fluctuation

PURPOSE: To determine the mean critical fusion frequency and the short-term fluctuation, to analyze the influence of age, gender, and the learning effect in healthy subjects undergoing flicker perimetry. METHODS: Study 1 - 95 healthy subjects underwent flicker perimetry once in one eye. Mean critical fusion frequency values were compared between genders, and the influence of age was evaluated using linear regression analysis. Study 2 - 20 healthy subjects underwent flicker perimetry 5 times in one eye. The first 3 sessions were separated by an interval of 1 to 30 days, whereas the last 3 sessions were performed within the same day. The first 3 sessions were used to investigate the presence of a learning effect, whereas the last 3 tests were used to calculate short-term fluctuation. RESULTS: Study 1 - Linear regression analysis demonstrated that mean global, foveal, central, and critical fusion frequency per quadrant significantly decreased with age (p<0.05).There were no statistically significant differences in mean critical fusion frequency values between males and females (p>0.05), with the exception of the central area and inferonasal quadrant (p=0.049 and p=0.011, respectively), where the values were lower in females. Study 2 - Mean global (p=0.014), central (p=0.008), and peripheral (p=0.03) critical fusion frequency were significantly lower in the first session compared to the second and third sessions. The mean global short-term fluctuation was 5.06±1.13 Hz, the mean interindividual and intraindividual variabilities were 11.2±2.8% and 6.4±1.5%, respectively. CONCLUSION: This study suggests that, in healthy subjects, critical fusion frequency decreases with age, that flicker perimetry is associated with a learning effect, and that a moderately high short-term fluctuation is expected.

Sensitivity and specificity of machine learning classifiers for glaucoma diagnosis using Spectral Domain OCT and standard automated perimetry

PURPOSE: To evaluate the sensitivity and specificity of machine learning classifiers (MLCs) for glaucoma diagnosis using Spectral Domain OCT (SD-OCT) and standard automated perimetry (SAP). METHODS: Observational cross-sectional study. Sixty two glaucoma patients and 48 healthy individuals were included. All patients underwent a complete ophthalmologic examination, achromatic standard automated perimetry (SAP) and retinal nerve fiber layer (RNFL) imaging with SD-OCT (Cirrus HD-OCT; Carl Zeiss Meditec Inc., Dublin, California). Receiver operating characteristic (ROC) curves were obtained for all SD-OCT parameters and global indices of SAP. Subsequently, the following MLCs were tested using parameters from the SD-OCT and SAP: Bagging (BAG), Naive-Bayes (NB), Multilayer Perceptron (MLP), Radial Basis Function (RBF), Random Forest (RAN), Ensemble Selection (ENS), Classification Tree (CTREE), Ada Boost M1(ADA),Support Vector Machine Linear (SVML) and Support Vector Machine Gaussian (SVMG). Areas under the receiver operating characteristic curves (aROC) obtained for isolated SAP and OCT parameters were compared with MLCs using OCT+SAP data. RESULTS: Combining OCT and SAP data, MLCs' aROCs varied from 0.777(CTREE) to 0.946 (RAN).The best OCT+SAP aROC obtained with RAN (0.946) was significantly larger the best single OCT parameter (p<0.05), but was not significantly different from the aROC obtained with the best single SAP parameter (p=0.19). CONCLUSION: Machine learning classifiers trained on OCT and SAP data can successfully discriminate between healthy and glaucomatous eyes. The combination of OCT and SAP measurements improved the diagnostic accuracy compared with OCT data alone.