京都大学 / 0048 / 新制・課程博士 / 博士(工学) / 甲第18303号 / 工博第3895号 / 新制||工||1598(附属図書館) / 31161 / 京都大学大学院工学研究科合成・生物化学専攻 / (主査)教授 濵地 格, 教授 森 泰生, 教授 跡見 晴幸 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DGAM

The focus of this study is the potential reduction of first language (L1) use during group activities in the second language (L2) classroom through the introduction of L2 support phrases with the goal of enabling learners to carry out three information exchange tasks without resorting to communication in L1. Many language instructors tend to limit group interaction because learners would typically revert to the language they are comfortable with and thus receive less comprehensive input and feedback from peers or get a chance to modify their output in the target language. Group work in the second language classroom is, however, most beneficial as it maximizes the number of turns students take and their overall speech production.  In the study I conducted, the participants were 16 students, forming eight pairs, from beginning Spanish language classes. Those in the L2 support class received linguistic assistance while those in the no L2 support class did not receive any additional help during the three information-exchange activities completed by the participants four to five weeks apart. The transcribed learner-learner interactions were analyzed in order to reveal the extent to which beginning Spanish language learners use the L1 rather than the target language during group work, and to make observations about the communicative functions of the languages used.  The findings revealed that providing L2 support phrases to manage tasks and deliberate over language use did not have an impact on the amount of the use of L1. On the contrary, it demonstrates that verbalizing less in the L1 in learner-learner interactions does not produce more target language. Thus, this study contributes to the theoretical explanation for the use of an L1 during learner-learner interactions, which does not always align with pedagogical practices. / Spanish

This project is aimed at generating new knowledge and improving our understanding of how Modified for Language Support-Worked Example Pairs (MLS-WEPs) contribute to effective mathematics learning and teaching in an ESOL (English to Speakers of Other Languages) context. The current study investigated a novel instructional approach to help English Language Learners (ELLs) develop better understanding in mathematical reasoning, problem solving, and literacy skills (listening, reading, writing, and speaking) while they are still developing their English language proficiency. The current study followed a wait-list control design, with both the treatment and control groups receiving intervention materials. The intervention materials were administered multiple times with different topics (units) throughout the study. The lessons were audio-recorded when the selected topics were taught. Pretest and posttest were given each time when the selected topics were taught. The data analysis for this study included both qualitative and quantitative analyses. The present study revealed the following results: (1) MLS-WEPs not only enhanced ELLs’ ability to solve mathematical problems, but also improved their written explanation skills and enabled them to transfer such skills to different mathematical concepts; (2) when controlling ELLs’ prior knowledge, the effectiveness of the MLS-WEPs intervention did not vary by their English language proficiency; (3) the MLS-WEPs intervention materials facilitated teachers to provide ELLs with more opportunities to read, write, and speak in mathematics and enabled teachers to ask more and deeper questions. However, worked example comparisons did not appear to motivate the participant teachers to promote equitable participation in mathematics classrooms. These findings provide direct empirical support for the need to reform mathematics teaching and learning in the ESOL context. / Math & Science Education

The goal of this project was to develop a lightweight Monte-Carlo (MC) event generator for hyperon pair production in antiproton-proton collisions, and demonstrate it in one of the hyperon decay reaction channels, which is collisions of proton and antiproton to lambda hyperon and anti-lambda hyperon, and they decay to antiproton-positive pion pair, and proton-negative pion pair. The prototype of this external MC event generator we presented here is based on the software framework of ROOT. Compared to the corresponding framework of embedded MC event generators that is currently used in the PANDA experiment at FAIR, PandaRoot, it allows for easier and quicker testing of new models or formalisms on synthetic data, for example for studying spin observables like polarisation. The event generator was benchmarked by studying angular distributions in final states, which are constructed by kinematic relations of 4-momentum vectors in different reference frames for all mother and daughter particles.

In this thesis I have theoretically studied electron recombination processes with small molecular ions. In these kind of processes resonant states are involved. To calculate the potential energy for these states as a function of internuclear distance, structure calculations and scattering calculations have to be performed. So far I have been studying the ion-pair formation with in electron recombination with H3+. The cross section for this process has been calculated using different kind of models, both a time dependent quantum mechanical and a semiclassical. I have also studied the direct process of dissociative recombination of HF+. To calculate the total cross section for this process, we have performed wave packet propagation on thirty resonant states and summed up the individual cross sections for these states. The cross sections for both these processes have a similar appearance to those measured experimentally in the ion storage ring CRYRING in Stockholm. / QC 20101103

Tidigare forskning visar att om lärare utformar NO-lektioner med inslag av kritiskt tänkande, skapar elever ett större positivt intresse till undervisningen och attityderna förändras till dess innehåll. Syftet med studien är att ta reda på hur NO-lärare arbetar med att utveckla elevers kritiska tänkande samt om de kan använda färdigt material. Semistrukturerade intervjuer har genomförts med sex olika mellanstadielärare och det visade sig att lärarna använder tre strategier för att nå dit, vilka är repetition, progression i olika NTA- laborationer samt skriva kritiska frågor och arbeta utefter EP A-modellen. Studien visar att de intervjuade lärarna anser att det är viktigt att utveckla elevers kritiska förmågor och lärarna arbetar aktivt med detta på lektionerna. Avslutningsvis framhåller studien att lärarna aktivt behöver arbeta med elevers kritiska tänkande och med NTA-materialet, samt att Concept Cartoons hjälper läraren i denna process. / Previous research shows that if teachers design their science classes with elements of critical thinking, pupils’ create a greater positive interest in the teaching and attitudes change towards its content. The purpose of this study is to find out how science teachers work to develop pupils' critical thinking and whether they can use ready-made material. Semi- structured interviews have been conducted with six different middle school teachers and it turned out that the teachers use three strategies to get there, which is repetition, progression in different NTA laboration and writing critical questions and working according to the Think- Pair-Share model. The study shows that the interviewed teachers believe it is important to develop pupils' critical abilities and the teachers work actively with this during class. In conclusion, the study highlights that teachers need to actively work with pupils' critical thinking and the NTA-material and Concept Cartoons help the teacher in this process.

The local structure, atomic and magnetic structure correlated over a small length scale, of a material has a strong impact on material properties. Pair distribution function (PDF) analysis is a strong tool to investigate local structure and magnetism of this nature. This work outlines extensions to current one dimensional magnetic pair distribution functions (1D-mPDF) and the fitting of structures with symmetry breaking local atomic distortions. 1DmPDF analysis has been used to study local magnetic structure, but requires a rotational averaging of the correlations so directional information is lost, as in powder diffraction experiments. Three dimensional difference magnetic pair distribution function (3D-∆mPDF) analysis does not require this rotational averaging and preserves directional information. This is a useful tool in analyzing experimental data like single crystal neutron diffraction and studying locally anisoptropic magnetic structures. Here we present a technique and software tools to calculate the 3D-∆mPDF pattern of a given structure and give a brief analysis of the local magnetic structure of MnTe. Another problem in PDF analysis is the modeling of structures with symmetry breaking local atomic distortions. Symmetry-adapted distortion modes have been used for structural refinement in Rietveld refinement for at least 10 years; more more recently, this has also been applied to PDF data. We present here a detailed discussion of the use of symmetry-adapted modes for structural refinement using PDF data. We also outline new open-sourced software tools to apply this technique and show two analyses using symmetry-adapted structural modes.

This bachelor's thesis in Computer Science explores the efficiency of an incremental re-tokenization algorithm in the context of BPE-trained SentencePiece models used in natural language processing. The thesis begins by underscoring the critical role of tokenization in NLP, particularly highlighting the complexities introduced by modifications in tokenized text. It then presents an incremental re-tokenization algorithm, detailing its development and evaluating its performance against a full text re-tokenization. Experimental results demonstrate that this incremental approach is more time-efficient than full re-tokenization, especially evident in large text datasets. This efficiency is attributed to the algorithm's localized re-tokenization strategy, which limits processing to text areas around modifications. The research concludes by suggesting that incremental re-tokenization could significantly enhance the responsiveness and resource efficiency of text-based applications, such as chatbots and virtual assistants. Future work may focus on predictive models to anticipate the impact of text changes on token stability and optimizing the algorithm for different text contexts.

One of the most important and challenging questions in the area of clustering is how to choose the best-fitting algorithm and parameterization to obtain an optimal clustering for the considered data. The clustering aggregation concept tries to bypass this problem by generating a set of separate, heterogeneous partitionings of the same data set, from which an aggregate clustering is derived. As of now, almost every existing aggregation approach combines given crisp clusterings on the basis of pair-wise similarities. In this paper, we regard an input set of soft clusterings and show that it contains additional information that is efficiently useable for the aggregation. Our approach introduces an expansion of mentioned pair-wise similarities, allowing control and adjustment of the aggregation process and its result. Our experiments show that our flexible approach offers adaptive results, improved identification of structures and high useability.

This essay investigates how the usage of minimal responses differs between men and women in different situations. There are three factors included in the analysis, gender of the interlocutor, formality, and the number of participants in the conversations. For each factor, the frequency and function of minimal responses are investigated. Eleven conversations are collected from the Santa Barbara Corpus to attain this aim. The method conducted to analyse the conversations is divided into two parts. The first part is to count the minimal responses used and calculate the frequency of usage for each speaker in all the conversations. The second part is a close analysis of the function of the minimal responses used by noticing whether they are disruptive or supportive. The results show that women use minimal responses at a higher frequency compared to men except in informal conversations. Also, there is no significant difference in the function of minimal responses between men and women. However, the minimal responses used in the informal conversations seem more disruptive. In pair and group conversations minimal responses can be used disruptively. However, if one considers the context, it seems that minimal responses in group conversations are collaborative, despite being disruptive.