A sharp increase in caesarean section (CS) rates at the Muhimbili National Referral Hospital (MNH) – a tertiary referral hospital in Tanzania – by 50% in 2000–2011, was associated with concomitant increase in maternal complications and deaths and inconsistent improvement in newborn outcomes. The aims of this thesis were to explore care providers’ in-depth perspective of the reasons for these high rates of CS, and to evaluate and improve standards of care for the most common indica-tions of CS, obstructed labour and fetal distress, which are also major causes of adverse maternal and neonatal outcomes. This thesis reports an investigation performed at MNH, Tanzania. For Paper I, qualitative methods were employed and demonstrated how care providers dismissed their responsibility for the rising CS rate; and, instead, projected the causes onto factors beyond their control. Additionally, dysfunctinal teamwork, transparency, and previous poorly conducted clinical audits led to fear of blame among care providers in cases of poor outcome that subsequently encougared defensive practise by assigning unnecessary CS. Papers II and III evaluated stand-ards of care using a criteria-based audit (CBA) of obstructed labour and fetal dis-tress. After implementing audit-feedback recommendations, the standards of diag-nosis of fetal distress improved by 16% and obstructed labour by 7%. Similarly, the standards of management preceding CS improved tenfold for fetal distress and doubled for obstructed labour. The impact of the CBA process was evaluated by comparing the maternal and perinatal outcomes categorized into Robson groups (Paper IV) of all deliveries occurring before and after the audit process (n=27,960). After the CBA process, there was a 50% risk reduction of severe perinatal morbidi-ty/mortality for patients with obstructed labour. The overall CS rates increased by 10%, and this was attributed to an increase in the CS rate among breech, term preg-nancies (Robson group 6), and preterm pregnancies (Robson group 10) that specifi-cally had reduced risk of poor perinatal outcome. The overall neonatal distress rates were also reduced by 20%, and this was attributed to a decrease in the neonatal distress rate among low-risk, term pregnancies (Robson group 3). Importantly, the increased rates of poor perinatal outcomes were associated with referred patients that had higher risk of neonatal distress and PMR than non–referred patients, after CBA process.  In conclusion, the studies managed to educate the care providers to take on their roles as decision-makers and medical experts to minimize unnecessary CS, using the available resources. Care providers’ commitment to achieve the best practice should be sustained and effort for stepwise upgrading quality of obstetric care should be supported by the hospital management from the primary to tertiary referral level.

Cervical cancer, caused by oncogenic (high risk [hr]) human papillomavirus (HPV) subtypes, is the most common cancer in women in Guatemala and the most common cause of cancer mortality in women aged 15-44 years. Visual inspection with acetic acid (VIA) with onsite cryotherapy “test-and-treat” is recommended for underserved Guatemalan indigenous rural women. This research assessed: 1) hrHPV infection prevalence in women screened by VIA; 2) Sensitivity and specificity of VIA in detecting hrHPV infection and cytologically identified precancerous and cancerous lesions; and 3) Factors associated with hrHPV infection. Analysis of anonymous data collected during VIA clinics in 2013 (N = 205) and 2017 (N = 234) for indigenous women aged 21-65 years in six villages showed 22.6% (95% confidence interval [CI]=18.7%-27.2%) had hrHPV cervical infection. VIA results were abnormal in 5.9% (95%CI=3.8%-8.8%). Only nine VIA exams in 89 women with hrHPV were abnormal (Sensitivity=10.1%, 95%CI=4.7%-18.3%), although abnormal VIA was associated with hrHPV (Prevalence Ratio [PR])=1.8; 95%CI=1.1-3.1; P=.05). Of 221 women who had VIA, hrHPV nucleic acid testing and liquid preparation cytology (equivalent to Papanicolaou or “Pap”) testing, 10 (4.7% [95%CI=2.3%-8.5%]) had abnormal cytological results, including one cancer, four high- and five low-grade squamous intraepithelial lesions. VIA sensitivity and specificity for detection of precancerous cytological abnormalities and cancerous lesions were 20.0% (95%CI=2.5%-55.6%) and 96.0% (95%CI=92.3%-98.3%) respectively. In contrast, hrHPV sensitivity and specificity were 100% (95%CI=71.7%-100%) and 88.7% (95%CI: 83.9%-92.7%). In both years combined, women aged fewer than 29 years or reporting fewer than four pregnancies were more likely to have hrHPV cervical infection (36.8%, 27.3%, respectively) than those who were older or reported more pregnancies (18.7; P=.025, respectively); 60.0% reported some form of modern contraception. Progesterone injections or implant users were more likely to have hrHPV infection (31.9%) than women using other or no contraceptives (19.5%); PR=1.6; 95%CI=1.1-2.4; P=.01). These data suggest that VIA may not be sufficiently sensitive for use in cervical cancer screening. “Test-and-treat” screening using hrHPV real-time testing, as recommended by the World Health Organization may be preferable to VIA, and may be acceptable using self-collected specimens.

The goal of this thesis is to create a dependency treebank for Yorùbá, a language with very little pre-existing machine-readable resources. The treebank follows the Universal Dependencies (UD) annotation standard, certain language-specific guidelines for Yorùbá were specified. Known techniques for porting resources from resource-rich languages were tested, in particular projection of annotation across parallel bilingual data. Manual annotation is not the main focus of this thesis; nevertheless, a small portion of the data was verified manually in order to evaluate the annotation quality. Also, a model was trained on the manual annotation using UDPipe.

<div>Structured prediction makes coherent decisions as structured objects to present the interrelations of these predicted variables. They have been widely used in many areas, such as bioinformatics, computer vision, speech recognition, and natural language processing. Machine Learning with reduced supervision aims to leverage the laborious and error-prone annotation effects and benefit the low-resource languages. In this dissertation we study structured prediction with reduced supervision for two sets of problems, sequence labeling and dependency parsing, both of which are representatives of structured prediction problems in NLP. We investigate three different approaches.</div><div> </div><div>The first approach is learning with modular architecture by task decomposition. By decomposing the labels into location sub-label and type sub-label, we designed neural modules to tackle these sub-labels respectively, with an additional module to infuse the information. The experiments on the benchmark datasets show the modular architecture outperforms existing models and can make use of partially labeled data together with fully labeled data to improve on the performance of using fully labeled data alone.</div><div><br></div><div>The second approach builds the neural CRF autoencoder (NCRFAE) model that combines a discriminative component and a generative component for semi-supervised sequence labeling. The model has a unified structure of shared parameters, using different loss functions for labeled and unlabeled data. We developed a variant of the EM algorithm for optimizing the model with tractable inference. The experiments on several languages in the POS tagging task show the model outperforms existing systems in both supervised and semi-supervised setup.</div><div><br></div><div>The third approach builds two models for semi-supervised dependency parsing, namely local autoencoding parser (LAP) and global autoencoding parser (GAP). LAP assumes the chain-structured sentence has a latent representation and uses this representation to construct the dependency tree, while GAP treats the dependency tree itself as a latent variable. Both models have unified structures for sentence with and without annotated parse tree. The experiments on several languages show both parsers can use unlabeled sentences to improve on the performance with labeled sentences alone, and LAP is faster while GAP outperforms existing models.</div>

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第21395号 / 情博第681号 / 新制||情||117(附属図書館) / 京都大学大学院情報学研究科社会情報学専攻 / (主査)教授 石田 亨, 教授 吉川 正俊, 教授 河原 達也 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

This research has been conducted in order to create a robust, lightweight feature detecting and matching algorithm that builds upon the foundation set by the TreeBASIS algorithm. The goal is to create a color-based version of the TreeBASIS algorithm that uses less hardware resources than the original, is more accurate in its matching capabilities, can successfully be deployed on a resource-limited FPGA platform, and can process in real time. This thesis first presents the newly designed hardware tri-channel FAST Feature Detector that finds features in color. Next the TreeBASIS algorithm is analyzed to discover what improvements can be made in order to reduce its resource usage sufficiently to be able to run on the Xilinx Virtex-4 FX60 while processing color features. At the same time, a software version of the Color TreeBASIS algorithm is compared to the original algorithm and is found to have a 93.3% accuracy on a test set of aerial images, surpassing the accuracy of TreeBASIS by nearly 12%. Then the hardware is meticulously reviewed to discover even more optimizations that allow the Color TreeBASIS algorithm to easily fit onto the Virtex-4 FX60. Next a new application for the matching algorithm, object detection, is introduced as well as the hardware needed to support it. Finally the algorithm is tested on the FPGA system for object detection and is able to successfully identify objects at 60 FPS. Color TreeBASIS proves itself to be more accurate than the TreeBASIS algorithm in the aerial images tests, it ends up using less memory and logic resources than its predecessor, even though it processes three times as much data, it is successfully deployed on a resource-limited FPGA system, and it shows accurate results in real-time object identification, generating an accurate homography 20 to 45% of the time while processing matches at a rate of 60 FPS.

No description available.

京都大学 / 新制・課程博士 / 博士(情報学) / 甲第25420号 / 情博第858号 / 新制||情||144(附属図書館) / 京都大学大学院情報学研究科知能情報学専攻 / (主査)特定教授 黒橋 禎夫, 教授 河原 達也, 教授 鹿島 久嗣 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

Video srauto šifravimas yra pagrindinis intelektinės nuosavybės užtikrinimo metodas šiuolaikinėse skaitmeninių teisių valdymo sistemose (DRM). Egzistuojantys šifravimo metodai užtikrina labai aukštą apsaugos lygį, bet tuo pat metu reikalauja labai didelių skaičiavimo resursų ir yra sunkiai pritaikomi ribotų skaičiavimo resursų mobiliems vartotojų įrenginiams. Šiame straipsnyje pristatomas dalinis MPEG-2 transporto srauto šifravimo metodas, kuris yra tinkamas tokiems ribotų resursų terminaliniams vartotojų įrenginiams. Siūlomo metodo eksperimentinio įvertinimo rezultatai rodo, kad šifravimo ir dešifravimo laikai, lyginant su standartiniu pilnu šifravimu, yra sumažinami apie 7 kartus, todėl šis metodas yra tinkamas ribotų skaičiavimo resursų dešifravimo įrenginiams. / Video stream encryption is the main method for the protection of intellectual property in modern digital rights management systems (DRM). Existing encryption methods ensure very high protection level of encrypted content, but at the same time are very resources demanding, making them hardly suitable for low processing power mobile end-user devices. This paper presents partial MPEG-2 transport stream encryption method, which is suitable for such low power user terminals. The experimental evaluation of proposed method show that encryption and decryption times, comparing with standard full encryption, are reduced about 7 times, making proposed method a good choice for low processing power decryption devices.

La diversité linguistique est actuellement menacée : la moitié des langues connues dans le monde pourraient disparaître d'ici la fin du siècle. Cette prise de conscience a inspiré de nombreuses initiatives dans le domaine de la linguistique documentaire au cours des deux dernières décennies, et 2019 a été proclamée Année internationale des langues autochtones par les Nations Unies, pour sensibiliser le public à cette question et encourager les initiatives de documentation et de préservation. Néanmoins, ce travail est coûteux en temps, et le nombre de linguistes de terrain, limité. Par conséquent, le domaine émergent de la documentation linguistique computationnelle (CLD) vise à favoriser le travail des linguistes à l'aide d'outils de traitement automatique. Le projet Breaking the Unwritten Language Barrier (BULB), par exemple, constitue l'un des efforts qui définissent ce nouveau domaine, et réunit des linguistes et des informaticiens. Cette thèse examine le problème particulier de la découverte de mots dans un flot non segmenté de caractères, ou de phonèmes, transcrits à partir du signal de parole dans un contexte de langues très peu dotées. Il s'agit principalement d'une procédure de segmentation, qui peut également être couplée à une procédure d'alignement lorsqu'une traduction est disponible. En utilisant deux corpus en langues bantoues correspondant à un scénario réaliste pour la linguistique documentaire, l'un en Mboshi (République du Congo) et l'autre en Myene (Gabon), nous comparons diverses méthodes monolingues et bilingues de découverte de mots sans supervision. Nous montrons ensuite que l'utilisation de connaissances linguistiques expertes au sein du formalisme des Adaptor Grammars peut grandement améliorer les résultats de la segmentation, et nous indiquons également des façons d'utiliser ce formalisme comme outil de décision pour le linguiste. Nous proposons aussi une variante tonale pour un algorithme de segmentation bayésien non-paramétrique, qui utilise un schéma de repli modifié pour capturer la structure tonale. Pour tirer parti de la supervision faible d'une traduction, nous proposons et étendons, enfin, une méthode de segmentation neuronale basée sur l'attention, et améliorons significativement la performance d'une méthode bilingue existante. / Language diversity is under considerable pressure: half of the world’s languages could disappear by the end of this century. This realization has sparked many initiatives in documentary linguistics in the past two decades, and 2019 has been proclaimed the International Year of Indigenous Languages by the United Nations, to raise public awareness of the issue and foster initiatives for language documentation and preservation. Yet documentation and preservation are time-consuming processes, and the supply of field linguists is limited. Consequently, the emerging field of computational language documentation (CLD) seeks to assist linguists in providing them with automatic processing tools. The Breaking the Unwritten Language Barrier (BULB) project, for instance, constitutes one of the efforts defining this new field, bringing together linguists and computer scientists. This thesis examines the particular problem of discovering words in an unsegmented stream of characters, or phonemes, transcribed from speech in a very-low-resource setting. This primarily involves a segmentation procedure, which can also be paired with an alignment procedure when a translation is available. Using two realistic Bantu corpora for language documentation, one in Mboshi (Republic of the Congo) and the other in Myene (Gabon), we benchmark various monolingual and bilingual unsupervised word discovery methods. We then show that using expert knowledge in the Adaptor Grammar framework can vastly improve segmentation results, and we indicate ways to use this framework as a decision tool for the linguist. We also propose a tonal variant for a strong nonparametric Bayesian segmentation algorithm, making use of a modified backoff scheme designed to capture tonal structure. To leverage the weak supervision given by a translation, we finally propose and extend an attention-based neural segmentation method, improving significantly the segmentation performance of an existing bilingual method.