This two-part dissertation has two goals: 1) a close philological reading of a 50-page section of a 10th-century Sanskrit philosophical work (Bhāsarvajña's Nyāyabhūṣaṇa), and 2) the creation and assessment of a novel intertextuality research system (Vātāyana) centered on the same work. The first half of the dissertation encompasses the philology project in four chapters: 1) background on the author, work, and key philosophical ideas in the passage; 2) descriptions of all known manuscript witnesses of this work and a new critical edition that substantially improves upon the editio princeps; 3) a word-for-word English translation richly annotated with both traditional explanatory material and novel digital links to not one but two interactive online research systems; and 4) a discussion of the Sanskrit author's dialectical strategy in the studied passage. The second half of the dissertation details the intertextuality research system in a further four chapters: 5) why it is needed and what can be learned from existing projects; 6) the creation of the system consisting of curated textual corpus, composite algorithm in natural language processing and information retrieval, and live web-app interface; 7) an evaluation of system performance measured against a small gold-standard dataset derived from traditional philological research; and 8) a discussion of the impact such new technology could have on humanistic research more broadly. System performance was assessed to be quite good, with a 'recall@5' of 80%, meaning that most previously known cases of mid-length quotation and even paraphrase could be automatically found and returned within the system's top five hits. Moreover, the system was also found to return a 34% surplus of additional significant parallels not found in the small benchmark. This assessment confirms that Vātāyana can be useful to researchers by aiding them in their collection and organization of intertextual observations, leaving them more time to focus on interpretation. Seventeen appendices illustrate both these efforts and a number of side projects, the latter of which span translation alignment, network visualization of an important database of South Asian prosopography (PANDiT), and a multi-functional Sanskrit text-processing web application (Skrutable).:Preface (i) Table of Contents (ii) Abbreviations (v) Terms and Symbols (v) Nyāyabhūṣaṇa Witnesses (v) Main Sanskrit Editions (vi) Introduction (vii) A Multi-Disciplinary Project in Intertextual Reading (vii) Main Object of Study: Nyāyabhūṣaṇa 104–154 (vii) Project Outline (ix) Part I: Close Reading (1) 1 Background (1) 1.1 Bhāsarvajña (1) 1.2 The Nyāyabhūṣaṇa (6) 1.2.1 Ts One of Several Commentaries on Bhāsarvajña's Nyāyasāra (6) 1.2.2 In Modern Scholarship, with Focus on NBhū 104–154 (8) 1.3 Philosophical Context (11) 1.3.1 Key Philosophical Concepts (12) 1.3.2 Intra-Textual Context within the Nyāyabhūṣaṇa (34) 1.3.3 Inter-Textual Context (36) 2 Edition of NBhū 104–154 (39) 2.1 Source Materials (39) 2.1.1 Edition of Yogīndrānanda 1968 (E) (40) 2.1.2 Manuscripts (P1, P2, V) (43) 2.1.3 Diplomatic Transcripts (59) 2.2 Notes on Using the Edition (60) 2.3 Critical Edition of NBhū 104–154 with Apparatuses (62) 3 Translation of NBhū 104–154 (108) 3.1 Notes on Translation Method (108) 3.2 Notes on Outline Headings (112) 3.3 Annotated Translation of NBhū 104–154 (114) 4 Discussion (216) 4.1 Internal Structure of NBhū 104–154 (216) 4.2 Critical Assessment of Bhāsarvajña's Argumentation (218)   Part II: Distant Reading with Digital Humanities (224) 5 Background in Intertextuality Detection (224) 5.1 Sanskrit Projects (225) 5.2 Non-Sanskrit Projects (228) 5.3 Operationalizing Intertextuality (233) 6 Building an Intertextuality Machine (239) 6.1 Corpus (Pramāṇa NLP) (239) 6.2 Algorithm (Vātāyana) (242) 6.3 User Interface (Vātāyana) (246) 7 Evaluating System Performance (255) 7.1 Previous Scholarship on NBhū 104–154 as Philological Benchmark (255) 7.2 System Performance Relative to Benchmark (257) 8 Discussion (262) Conclusion (266) Works Cited (269) Main Sanskrit Editions (269) Works Cited in Part I (271) Works Cited in Part II (281) Appendices (285) Appendix 1: Correspondence of Joshi 1986 to Yogīndrānanda 1968 (286) Appendix 1D: Full-Text Alignment of Joshi 1986 to Yogīndrānanda 1968 (287) Appendix 2: Prosopographical Relations Important for NBhū 104–154 (288) Appendix 2D: Command-Line Tool “Pandit Grapher” (290) Appendix 3: Previous Suggestions to Improve Text of NBhū 104–154 (291) Appendix 4D: Transcript and Collation Data for NBhū 104–154 (304) Appendix 5D: Command-Line Tool “cte2cex” for Transcript Data Conversion (305) Appendix 6D: Deployment of Brucheion for Interactive Transcript Data (306) Appendix 7: Highlighted Improvements to Text of NBhū 104–154 (307) Appendix 7D: Alternate Version of Edition With Highlighted Improvements (316) Appendix 8D: Digital Forms of Translation of NBhū 104–154 (317) Appendix 9: Analytic Outline of NBhū 104–154 by Shodo Yamakami (318) Appendix 10.1: New Analytic Outline of NBhū 104–154 (Overall) (324) Appendix 10.2: New Analytic Outline of NBhū 104–154 (Detailed) (325) Appendix 11D: Skrutable Text Processing Library and Web Application (328) Appendix 12D: Pramāṇa NLP Corpus, Metadata, and LDA Modeling Info (329) Appendix 13D: Vātāyana Intertextuality Research Web Application (330) Appendix 14: Sample of Yamakami Citation Benchmark for NBhū 104–154 (331) Appendix 14D: Full Yamakami Citation Benchmark for NBhū 104–154 (333) Appendix 15: Vātāyana Recall@5 Scores for NBhū 104–154 (334) Appendix 16: PVA, PVin, and PVSV Vātāyana Search Hits for Entire NBhū (338) Appendix 17: Sample Listing of Vātāyana Search Hits for Entire NBhū (349) Appendix 17D: Full Listing of Vātāyana Search Hits for Entire NBhū (355) Overview of Digital Appendices (356) Zusammenfassung (Thesen Zur Dissertation) (357) Summary of Results (361)

<p> A critical survey of the important features and characteristics of some existing Text-to-Speech Conversion (TSC) system by rules is given. The necessary algorithms, not available for these systems in the literature, have been formulated providing the basic philosophies underlying these systems. A new algorithm TESCON for a TSC system by rules is developed without implementation details. TESCON is primarily concerned with the preprocessing and linguistic analysis of an input text in English orthography. For the first time, the use of function-content word concepts is fully utilized to identify the potential head-words in phrases. Stress, duration modification and pause insertions are suggested as part of the rule schemes. TESCON is general in nature and is fully compatible with a true TSC system.</p> / Thesis / Master of Science (MSc)

<p> This work is an investigation of user acceptance of a prototype solution utilizing wireless text messaging (or SMS - i.e., short messaging service) to improve people's adherence. Insufficient adherence, also known as medical non-compliance, is a major cause of failure in self-management programs, causing significant losses to all healthcare stakeholders.</p> <p> Innovative mobile healthcare solutions, based on portable devices like cell phones, may address some non-adherence aspects by helping outpatients to follow treatments agreed with their health providers. Although this seems a win-win situation, a verdict on the overall usefulness of such an approach cannot be formulated before exploring outpatient acceptance, as this is a novel technology that targets a new area of implementation. Accordingly, this research investigates key factors that may influence the acceptance of a mobile healthcare solution based on SMS to support improved adherence to healthy behaviour, with special attention to motivation (the 'pro' factors) and perceived risk (the 'con' factors).</p> <p> As a means of investigation, a one-month longitudinal experiment with two groups of subjects (an intervention group and a control group) was utilized. Data were analyzed with quantitative and qualitative techniques: descriptive statistics, partial least squares modelling, and content analysis.</p> <p> Findings show that users are aware of the potential usefulness of such a pioneering application. However, enjoyment is the unique reason for adopting, and perceived financial and psychological risks the main obstacles against adopting, an SMS-based solution for improving adherence to healthy behaviour. Furthermore, a business analysis shows that users are concerned about usefulness features, even when asked about financial aspects.</p> <p> These results, together with encouraging findings about the effectiveness of the application, open the way for medical-led research to investigate if long-term mobile healthcare initiatives customized to patient needs are also beneficial for outpatient adherence and health outcomes.</p> / Thesis / Doctor of Philosophy (PhD)

京都大学 / 新制・課程博士 / 博士(情報学) / 甲第24928号 / 情博第839号 / 新制||情||141(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)教授 森 信介, 教授 西野 恒, 教授 中村 裕一 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

The rapid growth of user-generated content on social media platforms in the form of text caused moderating toxic language manually to become an increasingly challenging task. Consequently, researchers have turned to artificial intelligence (AI) and machine learning (ML) models to detect and classify toxic comments automatically. However, these models often exhibit unintended bias against comments containing sensitive terms related to de- mographic groups, such as race and gender, which leads to unfair classifications of samples. In addition, most existing research on this topic focuses on fully supervised learning frame- works. Therefore, there is a growing need to explore fairness in semi-supervised toxicity detection due to the difficulty of annotating large amounts of data. In this thesis, we aim to address this gap by developing a fair generative-based semi-supervised framework for mitigating social bias in toxicity text classification. This framework consists of two parts, first, we trained a semi-supervised generative-based text classification model on the bench- mark toxicity datasets. Then, in the second step, we mitigated social bias in the trained classifier in step 1 using adversarial debiasing, to improve fairness. In this work, we use two different semi-supervised generative-based text classification models, NDAGAN and GANBERT (the difference between them is that the former adds negative data augmenta- tion to address some of the problems in GANBERT), to propose two fair semi-supervised models called FairNDAGAN and FairGANBERT. Finally, we compare the performance of the proposed fair semi-supervised models in terms of accuracy and fairness (equalized odds difference) against baselines to clarify the challenges of social fairness in semi-supervised toxicity text classification for the first time. Based on the experimental results, the key contributions of this research are: first, we propose a novel fair semi-supervised generative-based framework for fair toxicity text classification for the first time. Second, we show that we can achieve fairness in semi- supervised toxicity text classification without considerable loss of accuracy. Third, we demonstrate that achieving fairness at the coarse-grained level improves fairness at the fine-grained level but does not always guarantee it. Fourth, we justify the impact of the labeled and unlabeled data in terms of fairness and accuracy in the studied semi- supervised framework. Finally, we demonstrate the susceptibility of the supervised and semi-supervised models against data imbalance in terms of accuracy and fairness.

No description available.

No description available.

No description available.

No description available.

No description available.