El presente documento evalúa el proceso de inversión de las administradoras de fondos de pensiones (AFP) en el Perú durante el periodo comprendido entre el 2007 y el 2015, inclusive. De este modo, busca proveer evidencia empírica referente a que hubiese sido posible obtener mayores rendimientos en los fondos que los observados históricamente por las AFP, de no existir restricciones como los límites a invertir en el extranjero o las ventas en corto. En este proceso, se analiza el portafolio de mercado a nivel del sistema para el fondo 2, por ser este el más representativo, y se determinan las clases de activos que lo constituyen, para luego, utilizando optimización inversa, hallar los retornos implícitos de estos. Asimismo, a través del momentum de las series de activos, con tres periodos de rezago, se especifican las opiniones o predicciones sobre el retorno de los mismos para luego ser combinadas con los retornos implícitos a través del modelo Black- Litterman. Así, se obtienen nuevos pesos de las clases de activos para cada uno de los meses de estudio y se calcula la rentabilidad del portafolio Black-Litterman acumulada, la cual resulta siendo considerablemente mayor a la rentabilidad acumulada histórica del sistema. / Tesis

This dissertation was conceived as an addendum to two self-published catalogs that American artist Chris Burden released, covering the years 1971–1977. It looks in-depth at the formative work the artist produced in college and graduate school, including minimalist sculpture, interactive environments, and performance art. Burden’s work is herewith examined in four chapters, each of which treats one or more related works, dividing the artist’s early career into developmental stages. In light of a wealth of new information about Burden and the atmosphere in which he was working in the late 1960s and early 1970s, this dissertation examines the artist’s work in relation to West Coast Minimalism, the Light and Space Movement, Environments, and Institutional Critique, above and beyond his well-known contribution to performance art, which is also covered herein. The dissertation also analyzes the social contexts in which Burden worked as having informed his practice, from the beaches of Southern California, to rock festivals and student protest on campus, and eventually out to the countercultural communes. The studies contained in the individual chapters demonstrate that close readings of Burden’s work can open up to formal and art-historical trends, as well as social issues that can deepen our understanding of these and later works. Benefitting from access to the artist’s estate, as well as archives collected at various institutions in Southern California, this dissertation is the first authoritative coverage of Chris Burden’s early career.

Database outsourcing is an emerging paradigm for data management in which data are stored in third-party servers. With the advance of cloud computing, database outsourcing has become popular and highly adopted. However, as a result, many technology challenges have arisen. / In this thesis, we study two problems with respect to the challenges, and propose solutions for each problem with the consideration of access patterns. The first problem is raised from theviewpoint of service providers. We study the problem of data allocation in scalable distributed database systems for achieving the high availability feature of cloud services. We propose a data allocation algorithm, which makes use of time series models from previous access patterns to perform load forecasting and reallocate data fragments to balance the workload within the system. Simulation results show that, with accurate forecasting, the proposed algorithm gives a better performance than general threshold-based algorithms. / The second problem addresses the clients' concern that service providers may not be trustworthy. We first illustrate how service providers can infer sensitive information through query access patterns even when data are encrypted. Then, we propose techniques that break down large queries and randomize query access patterns such that service providers cannot infer sensitive information with a high degree of certainty. Experiments on benchmark data show that a high level of access privacy can be achieved by the proposed techniques with a reasonable overhead. / 數據庫外包是近年新興的一種數據管理服務，其特點是數據儲存於第三方的伺服器內。隨著雲端科技的發展，數據庫外包服務日趨普及，同時亦產生不少技術問題。 / 本文著重探討兩個問題。首先，從服務供應商的角度研究可擴展的分布式數據庫系統如何分配數據來提供高可用性的雲端服務。鑑於用戶訪問模式會隨著時間轉變，我們提出以時間序列模型預測負荷的算法重新分配數據，以平衡系統的工作量。通過模擬實驗可知在準確的負荷預測下，我們提出的算法比基於闆值的算法有更好的表現。 / 第二個探討的問題是如何保障用戶私隱，避免洩露給服務供應商。文中列舉了數據加密的情況下，服務供應商如何通過分析用戶訪問模式獲取資料，進而提出相應的保障技術。透過用戶訪問模式的隨機化，能使服務供應商無法準確比對用戶的資料。基準數據實驗指出此項技術可有效保護私隱，而且不會對訪問速度造成太大影響。 / Li, Shun Pun. / Thesis M.Phil. Chinese University of Hong Kong 2015. / Includes bibliographical references (leaves 86-93). / Abstracts also in Chinese. / Title from PDF title page (viewed on 11, October, 2016). / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only.

As the bandwidth has become much cheaper in the recent years, video applications are more popular than before. However, the demand of high video resolution, high frame rate, or high bit-depth has continued to increase more rapidly than the cost of video transmission and storage bandwidth. It requires more efficient compression techniques, and hence many international video coding standards are developed in the past decades such as the MPEG-1/2/4 part 2, H.264/MPEG-4 part 10 AVC and the latest High Efficiency Video Coding (HEVC) standards. The main objective of this thesis is to consider the problems in analyzing the characteristics of video signals and providing efficient compression and transmission solutions in both H.264/AVC and HEVC video systems. Three main parts of this work are briey summarized below. / The first part concerns transform coding. Transform coding has been widely used to remove spatial redundancy of prediction residuals in the modern video coding standards. However, since the residual blocks exhibit diverse characteristics in a video sequence, conventional sinusoidal transforms with fixed transform kernels may result in low coding efficiency. To tackle this problem, we propose a novel content adaptive transform framework for H.264/AVC-based video coding. We propose to utilize pixel rearrangement to dynamically adjust the transform kernels to adapt to the video signals. In addition, unlike the traditional adaptive transforms, the proposed method obtains the transform kernels from the reconstructed block, and hence it consumes only one logic indicator for each transform unit. Moreover, a spiral-scanning method is developed to reorder the transform coefficients for better entropy coding. Experimental results on the Key Technical Area (KTA) platform show that the proposed method can achieve a significant bits reduction under both the all-intra and low-delay configurations. / The second part investigates the next-generation video coding. Due to increase of display resolution from High-definition (HD) to Ultra-HD, how to efficiently compress the Ultra-HD signals are essential in the development of future video compression systems. High-resolution video coding benefits from a larger prediction block size and thereof transform and quantization of prediction residues. However, in the current HEVC video coding standard, the maximum coding tree unit (CTU) size is 64x64, which can limit a possible larger prediction block in Ultra-HD video coding, and hence cause negative effects on coding efficiency. Thus, we propose to extend CTU to a super coding unit (SCU) for next-generation video coding, and two separate coding structures are designed to encode a SCU, including Direct-CTU and SCU-to-CTU modes. In Direct-CTU, an SCU is first split into a number of predefined CTUs, and then, the best encoding parameters are searched from the current CTU to the possible minimum coding unit (MCU). Similarly, in SCU-to-CTU, the best encoding parameters are searched from SCU to CTU. In addition, the adaptive loop filter (ALF) and sample adaptive offset (SAO) methods are investigated in SCU based video coding framework. We propose to change the filtering control from SCU level to the coding unit (CU) level, and an improved CU level ALF signaling method is also proposed to further improve the coding efficiency. Furthermore, an adaptive SAO block method is also proposed, and this flexibility of SAO blocks can further improve the performance of the traditional method in the Ultra HD video coding. / In the last part, we explore the bit rate control of video transmission. Rate control serves as an important technique to regulate the bit rate of video transmission over a limited bandwidth and to maximize the overall video quality. Video quality fluctuation plays a key role in human visual perception, and hence many rate control algorithms have been widely developed to maintain a consistent quality for video communication. We propose a novel rate control framework based on the Lagrange multiplier in HEVC. With the assumption of constant quality control, a new relationship between the distortion and the Lagrange multiplier is established. Based on the proposed distortion model and buffer status, we obtain a computationally feasible solution to the problem of minimizing the distortion variation across video frames at the coding tree unit level. Extensive simulation results show that our method outperforms the HEVC rate control by providing a more accurate rate regulation, lower video quality fluctuation and stabler buffer fullness. / 近些年，隨著帶寬費用變得越來越便宜，各種視頻應用比以前更為流行了。然而，人們對于高視頻分辨率，高幀率，或更高比特像素的需求增加了視頻傳輸和存儲帶寬的成本。滿足這樣的需求需要更有效的壓縮技術，因此在過去的幾十年裏，很多國際視頻編碼標准被開發出來，例如MPEG-1/2/4 part2, H264/MPEG-4 part 10 AVC和最新高效視頻編碼標准(HEVC)。本論文的主要目的是研究視頻信號的特點，在H.264和HEVC視頻系統中提供高效的壓縮和傳輸解決方案。論文分三部分，簡要總結如下。 / 第壹部分涉及變換編碼。在現代視頻編碼標准中，變換編碼已被廣泛用于消除預測殘差的空間冗余度。然而，由于在視頻序列中的預測殘差塊有著不同的特性，傳統的變換采用固定變換矩陣可能會導致低的編碼效率。為了解決這個問題，我們提出了壺種新的基于內容自適應變換方案的視頻編碼框架。我們利用重排像素，動態調整的變換矩陣以適應當前的視頻信號。此外，與傳統的自適應變換不同之處在于，我們所提出的方法得到的變換矩陣不需要傳輸到解碼端，而它僅消耗壺個邏輯單元指示當前變換矩陣。此外，我們提出了相應的變換系數掃描方法以達到更有效的熵編碼。在關鍵技術領域(KTA)平台，實驗結果表明本方法可以有效的改善幀內和低延遲的配置下的編碼效率。 / 第二部分探討了新壹代視頻編碼。由于主流顯示分辨率從高清到超高清的變化，如何有效地壓縮超高清視頻信號是未來視頻壓縮技術發展的關鍵。超高分辨率視頻編碼的好處在于可從壹個更大的預測塊對其預測殘差進行變換和量化。然而，在目前HEVC視頻編碼標準，最大編碼榭單元尺寸（CTU）是64x64，其可能限制較大的預測塊，從而影響編碼效率。因此，我們提出了擴展CTU為SCU。其中編碼壹個SCU可能用到兩個獨立的編碼模式，包括Direct-CTU和SCU-to-CTU。在Direct-CTU模式中，SCU被分割成許多預定義的CTUs，然後，最佳的編碼參數搜索範圍為CTU到MCU。同樣，在SCU-to-CTU模式中，最佳的編碼參數搜索範圍是SCU到CTU。此外，自適應環路濾波器（ALF）和自適應采偏移（SAO）在新的SCU編碼框架下進行了研究。我們提出濾波控制從SCU級別更改為CU級別，並提出了新的ALF信號傳送方法進壹步提高傳統的方法在超高清視頻編碼的中性能。 / 在最後壹部分，我們探討了視頻傳輸中的碼率控制。碼率控制作為壹種重要的技術，在有限的帶寬條件下，以最大限度地提高整體的視頻質量。視頻質量波動在人眼視覺感知中起著至關重要的作用，因此許多碼率控制方法得到了廣泛的發展，以追求提供穩定的視頻通信質量。我們提出了壹個新基于HEVC拉格日乘數碼率控制框架。在平穩視頻質量的假設下，我們提出了壹種新的失真和拉格日乘子之間的關係。基于新提出的失真模型和緩沖區的狀態，我們得到壹個計算上可行的解決方案，以最大限度地減少在編碼榭單元級的視頻幀的失真變化。大量的仿真結果表明，我們的方法優于HEVC的碼率控制，它可以提供更精確的碼率調節，降低視頻質量波動，以及維護穩定的緩沖區占有率。 / Wang, Miaohui. / Thesis Ph.D. Chinese University of Hong Kong 2015. / Includes bibliographical references (leaves 158-164). / Abstracts and acknowledgements also in Chinese. / Title from PDF title page (viewed on 11, October, 2016). / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only.

Background. Mounting research has established child maltreatment as a significant risk factor for negative mental health outcomes in adulthood. However, the mechanisms through which childhood maltreatment exerts its detrimental effect on psychological functioning in adulthood are not well understood. Attachment theory and cognitive theory provide useful theoretical frameworks in explaining the etiology of adult psychopathology, but no prior research has tested the mediating mechanisms of insecure attachment pattern, intolerance of uncertainty (IU) and rumination within a unified model. / Objectives. There are three main objectives in the current study: (1) to explore the relationships among child maltreatment history and adult depressive and anxiety symptoms, as well as insecure attachment pattern, IU and rumination; (2) to test a etiological mediation model with distal child maltreatment predicting adult emotional outcomes via the mediating processes of insecure attachment pattern, IU and rumination; and (3) to investigate whether rumination would mediate the path between IU and adult emotional outcomes. / Method. A total of 322 adults receiving clinical psychological service in the Clinical Psychology Units of the Social Welfare Department, Hong Kong Special Administrative Region, completed the questionnaires with measures on early childhood maltreatment, attachment pattern, IU, rumination, adult emotional outcomes and proximal life stress. Structural equation modeling was conducted for data analysis. / Results. The present study showed that child maltreatment was incredibly common among clinical population and it significantly affected adult emotional outcomes as well as the three mediators (insecure attachment pattern, IU and rumination) examined. A Developmental IU Rumination-Attachment Model with good model fit and strong explanatory power was established. Two sets of etiological paths: Rumination Paths (Simple Rumination Indirect Path and IU-mediated Rumination Path) and Attachment Paths (Simple Attachment Indirect Path and IU-mediated Attachment Path) were identified. The model revealed a novel finding that IU was a common vulnerability factor for both insecure attachment pattern and rumination. Rooted from IU, insecure attachment pattern exemplified stronger influence on adult emotional outcomes when comparing with ruminative response style. Ruminative response style mediated the development from IU to adult emotional outcomes. Proximal life stress cast the least influence on adult emotional outcomes when comparing with the mediating mechanism examined. / Discussion. The Developmental IU Rumination-Attachment Model suggests that in the process of developing adult emotional problems, most of the influences stemmed from pre-existing child maltreatment history and internal psychological processes involving insecure attachment pattern, IU and ruminative response style. External proximal negative life events contribute only a relatively small proportion of influence. The current model does not only enrich our understanding to the long-term sequelae of child maltreatment, but also helps to inform important treatment components for survivors of child maltreatment and to shed light on preventive measures for child maltreatment. Implications for future research would also be discussed. / 背景：大量文獻早已確立孩童期受虐經歷是成人身心健康發展的高危因素。然而，探討孩童期受虐經歷如何影響受虐者長大後的情緒健康的研究，到目前為止卻是不多。雖然依附理論及認知理論對成人心理問題的發展進程奠定了重要的理論基礎，然而至今仍沒有任何研究同時探討當中的三個重要中介變量：依附模式，無法忍受不確定性及反芻。 / 研究目的：本研究有三個主要目的：(1) 探討孩童期受虐經歷與成人抑鬱及焦慮情緒、依附模式、無法忍受不確定性及反芻之間的關係；(2) 測試一個層次性病原模型，闡述孩童期被虐經歷的遠端因素，如何通過一系列的中介變量（依附模式、無法忍受不確定性及反芻），發展至成人期的情緒問題；(3) 研究反芻能否被確立為無法忍受不確性與成人情緒問題的中介變量。 / 研究方法：研究以問卷調查方式（調查內容包括孩童期受虐經歷、依附模式、無法忍受不確定性、反芻、成人情緒健康及近端生活壓力），於322 位在香港特別行政區社會福利署臨床心理服務課接受服務的人士所組成的樣本，使用結構方程模型進行數據分析。 / 研究結果：結果顯示，孩童期受虐經歷在臨床群體中是非常普遍，而孩童期受虐經歷也顯著地影響成人期的抑鬱及焦慮情緒，以及三個被測試的中介變量。研究所得的「發展性無法忍受不確定性反芻依附模型」（Developmental IURumination-Attachment Model）達到良好的摸型擬合度及豐富的臨床意義。模型包含了兩組病原路徑，分別為反芻路徑（單一反芻間接路徑和無法忍受不確定性中介的反芻路徑）及依附路徑（單一依附間接路徑和無法忍受不確定性中介的依附路徑）。此模型揭示了一個斬新的發現：無法忍受不確定性乃是誘發反芻及不安全依附模式的脆弱性因素。在無法忍受不確定性的情況下，依附模式較反芻對成人期的情緒問題更具影響力。除此以外，反芻被確立為無法忍受不確定性與成人情緒問題的中介變量。相對模型中的三個中介變量，近端生活壓力對成人情緒問題所構成的影響力則最弱。 / 討論：研究所得的模型有效地揭示了成人期情緒問題的發展過程中，近端外在生活壓力只構成較弱的影響力。遠端孩童期的受虐經歷，至一系列較近期的內在心理現象（依附模式、無法忍受不確定性及反芻）才是最為重要的因素。研究結果不但加深了學術界對兒童期受虐經歷長遠影響的認識，亦為預防及治療成人情緒問題提供多個臨床應用策略。最後，文未也就未來的實證研究提出了建議。 / Lam, Shui Mun Tracy. / Thesis Ph.D. Chinese University of Hong Kong 2015. / Includes bibliographical references (leaves 180-214). / Abstracts also in Chinese; appendixes in Chinese. / Title from PDF title page (viewed on 11, October, 2016). / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only.

Crustaceans are ubiquitous in the world’s oceans and include species ranged in size from < 1 mm to > 1 m. Information on the natural diet of marine crustaceans, especially the small crustaceans such as copepods, is usually obtained by measuring gut content microscopically or by the use of electronic particle counters, by microscopic examination of gut contents and fecal pellets, and by the use of stable isotopes or taxon-specific pigment markers. Unfortunately, most of these methods involve conducting feeding experiments in artificial laboratory condition, require a high level of taxonomic expertise, and provide only coarse information on food types that are digestion-resistant. Increasingly over the past decade, DNA-based methods have been used to study the natural diet of small marine organisms such as molluscan larva and the lobster larva with high taxonomic resolution. / In this thesis, the diets of two marine crustaceans are investigated. The small calanoid copepod Calanus sinicus (2.5−3.5 mm), one of the animals studied in this thesis, is a dominant member of the marine zooplankton in the continental shelf waters of China and the northwest Pacific Ocean. Copepods of the genus Calanus are generally considered to be herbivorous although information on the natural diet of C. sinicus is scarce. / Another marine crustacean studied in this thesis is the hydrothermal crab Xenograpsus testudinatus. This crab is endemic to the shallow-water hydrothermal vents of Kueishan Island in the northeastern corner of Taiwan. The extreme environment inhabited by X. testudinatus is unsuitable for most marine organisms due to discharge of hot and acidic water with high content of sulphur. Although the diet of X. testudinatus has been studied by several investigations, the taxonomic resolution of the dietary information is poor. / Accurate information of the diet and feeding habit of a species provides the basis for understanding its role in aquatic food webs. In my thesis, the natural diets of the marine copepod C. sinicus from Hong Kong and Taiwan and the hydrothermal vent crab X. testudinatus were investigated using Illumina sequencing of the prey DNA in the gut content. Variable regions of the 18S and 16S rRNA genes in the gut contents are amplified using DNA extracted from gut contents to determine the composition of eukaryotes and prokaryotes in the diet of the marine crustacean. The diet of X. testudinatus are compared to those of the mangrove crabs Perisesarma bidens. / Gelatinous zooplankton including Hydrozoa and unclassified Ctenophora are very common in the gut contents of C. sinicus from both Hong Kong and Taiwan. C. sinicus from Hong Kong contains diatoms and dinoflagellates, while C. sinicus from northern Taiwan also feeds on Anthozoa, Ascidiacea and Malacostraca. Unclassified Proteobacteria and unclassified bacteria are common in C. sinicus from both Hong Kong and Taiwan. Cyanobacteria, mainly Synechococcus, are only found in C. sinicus from Taiwan. My results show that C. sinicus can feed omnivorously on a diverse assemblage of phytoplankton and zooplankton. / The diet of X. testudinatus differs from that of P. bidens. While P. bidens feeds mainly on Magnoliopsida plant, X. testudinatus feeds on a diverse assemblage of organisms including algae, fishes, bivalves, copepods and anthozoans. Although soil bacteria such as Rodobacteraceae and cyanobacteria including Oscillatoriphycidae occur in the gut contents of both crabs, Mycoplasmataceae and Helicobacteraceae, which are suspected to be gut probiotic bacteria, are only found in the vent crabs. My findings suggest that both X. testudinatus and P. bidens are scavengers. / In short, the DNA-based method can provide detailed information on the natural diets of marine crustaceans and provide accurate information on the trophic role of the marine crustaceans in the food webs. / 甲殻類動物在海洋上是很常見的。牠們的體型大小也有很大的差異(由< 1毫米到 > 1米)。有關海洋甲殻類動物食性研究特別是一些細小甲殻類動物例如橈腳類動物通常都是利用電子計量器或以肉眼在顯微鏡下計算，利用電子顯微鏡去鑑定腸及糞便中的食物殘滓及利用穩定同位素或物種獨有的色素去鑑定。不過，這些方法包括實驗室餵食測試需要熟練的鑑定技巧而且這些方法只能夠提供粗略的資訊。例如只限於會經消化後留下硬塊的食物。雖然如此，最近有研究成功利用DNA鑑定的方法去找出一些細小海洋甲殻類動物例如龍蝦及軟體動物的幼體的食性。 / 這篇論文主要研究兩種海洋甲殻類動物的野外食物。其中一種研究對象，細小的橈腳類動物中華哲水蚤 (2.5−3.5毫米)，在中國沿岸海域及西北太平洋是很常見的。雖然有關於中華哲水蚤在食性資料不多，但是哲水蚤類的動物一般被認為是草食性。 / 這篇論文的另一研究對象是烏龜怪方蟹。牠在位於台灣東北部的龜山島的淺水熱噴泉是很常見的。在淺水熱噴泉的生境，由於經常有酸性高溫及含有高濃度硫磺的水噴出，所以那裡是被認為不適合生物居住的。雖然烏龜怪方蟹的食性也有被研究過，但有關牠們的攝食的物種鑑定還是不夠詳細。 / 準確的有關生物的食性的資料可以幫助我們了解該生物在海洋食物網中角色。有見及此，在這篇論文中主要會利用DNA鑑定及Illumina定序技術來研究來自香港及台灣的中華哲水蚤及在淺水熱噴泉的烏龜怪方蟹的食性。在實驗中，腸道中的DNA的18S及16S rRNA基因中DNA序列差異大的區域會被用來鑑定腸道中的有核生物及原生生物。居住在紅樹林的雙齒近相手蟹的攝食也會用來跟烏龜怪方蟹比較。 / 啫哩狀的浮游動物比如水息蟲類動物及櫛水母來自香港及台灣的中華哲水蚤的腸道中是很常見的。在來自香港的中華哲水蚤腸道中找到了硅藻及甲藻，而在來自台灣的中華哲水蚤的腸道中卻找到珊瑚綱動物，海蛸和軟甲綱動物。在來自香港及台灣的中華哲水蚤的腸道中找到的原核生物為不能鑑定的細菌及不能鑑定的變形菌。藍綠藻主要為聚球藻只有在來自台灣的中華哲水蚤的腸道中找到。研究的結果顯示中華哲水蚤是雜食性及攝食很多不同類型的浮游生物。 / 研究結果顯示烏龜怪方蟹所食的食物跟雙齒近相手蟹很不同。雙齒近相手蟹主要攝食雙子葉植物綱植物，而烏龜怪方蟹所吃的食物是高生物多樣性的包括藻類，魚，雙殻綱動物，橈腳類動物及珊瑚綱動物。土壤細菌比如紅桿菌科細菌及藍綠藻都可以烏龜怪方蟹及雙齒近相手蟹的腸道及胃道中找到，而可能是腸道原生細菌的枝原體科細菌及螺旋杆菌科細菌只能夠在烏龜怪方蟹的腸道中找到。研究結果顯示烏龜怪方蟹及雙齒近相手蟹都是拾荒者。 / 總括而言，利用DNA去研究動物食性的方法可以提供更多有關海洋甲殻類動物食性的詳細資料及提供準確的有關海洋甲殻類動物食性層次的資料。 / Ho, Tsz Wai. / Thesis M.Phil. Chinese University of Hong Kong 2015. / Includes bibliographical references (leaves 65-85). / Abstracts also in Chinese. / Title from PDF title page (viewed on 12, October, 2016). / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only.

The complexity of English prepositions has always been a challenge for second language (L2) learners (Celce-Muricia & Larsen-Freeman, 1999). Cognitive linguists have made a substantial amount of effort of analyzing the semantic representations of the prepositions (Lakoff, 1987; Talmy, 2000; Tyler & Evans, 2003; Lindstromberg, 2010). Schematic drawings, with variation among studies, are proposed to visualize the conceptual representations and to connect the basic spatial senses (e.g., in a box) with the more abstract, metaphorical senses (e.g., in love) of prepositions. Some empirical studies (Tyler, 2012; Tyler, Mueller & Ho, 2010) found positive learning outcomes when L2 learners were taught prepositions with cognitive linguistics methods. Yet, few studies provided step-by-step descriptions of their instructional procedures. Inadequate information of exactly what was going on in these classrooms makes it difficult to replicate or to compare across studies. / The current study is an experimental study that investigates the effects of teaching English prepositions with cognitive linguistics methods in a computer-based tutorial system. Our main research question is to see whether cognitive linguistic approach outperforms traditional explicit instruction. The participants of the present study were 64 Cantonese learners of English, at Secondary 4 level, from a Hong Kong secondary school with English as the medium of instruction. They were randomly assigned to one of the four groups: 1) schema feedback (SF); 2) rule feedback (RF); 3) corrective feedback (CF); and 4) control group (receiving training on English articles). The first three groups received training on three prepositions (in, at, over) through the online interface shown in Figure 1. On each screen they were presented with a picture stimulus and two minimal-pair sentences that only differed on the prepositions (e.g., Ben jumped over the ditch vs. Ben jumped towards the ditch). They were asked to choose the sentence that matched the picture. They were provided with immediate feedback. The SF group (Figure 1) was given the schematic drawing mapped with the sentence as well as a verbal explanation of the mapping. The RF group was given a metalinguistic explanation of the preposition usage with three example sentences. The CF group was informed of whether their choice was correct or wrong. All the four groups went through the same procedure with three sessions: a pretest, training (balanced time across groups), and a posttest. The tests were composed of two parts: a cloze test and a translation test, both at the sentential level. / The findings showed no significant difference in terms of accuracy of performance among the four groups at the pretest. But at the posttest, there was a significant difference between the three preposition training groups and the control group. This finding suggests positive instructional effects of the tutor. Regarding the distinction of feedback, a marginal significant difference was observed among the SF, RF, and CF groups at the posttests and such difference only came from the translation test. That indicates that the amount and the type of information that were provided to the participants did not matter for cloze test but did create superiority effect in the production test, when they were instructed with a minimal-pair paradigm and also with the help of the picture stimuli.[with diagram] / 英語介詞，因其複雜性，對於第二語言（L2）學習者一直是學習上的一個挑戰（Celce-Muricia & Larsen-Freeman, 1999）。認知語言學家一直努力分析介詞的語義表述（Lakoff, 1987; Talmy, 2000; Tyler & Evans, 2003; Lindstromberg, 2010）。有研究提出以不同的圖像模組將概念表述形像化及連接基本的空間意義（例如，in a box）及較抽象的隱喻意義（例如，in love）。一些實證研究（Tyler, 2012; Tyler, Mueller & Ho, 2010）中，發現第二語言學習者以認知語言學的方法學習介詞時，學習成果會較好。然而，很少有研究詳細地描述自己的教學過程，因此資訊的不足令它其難以模仿或比較不同的研究。 / 目前的研究是一個實驗性的研究以探討在電腦系統上以認知語言學的方法教導介詞的效果。主要研究的問題是，究竟認知語言學的方法（圖像模組）能否超越傳統教學方法（給予明確指引）。研究的參加者是64位以廣東話作為母語，在香港中學就讀中四的英語學習者。他們被隨機分配到四個組別中的其中一個組別：1）圖像模組的反饋（SF）;2）規則和例子的反饋（RF）;3）糾正反饋（CF）和4）對照組（接受英語冠詞的訓練）。前三組會透過線上練習（圖1）接受3個介詞（in, at, over）的訓練。每個練習上會有一幅圖片及一組對比句子（兩句句子不同之處只在於介詞，例如Ben jumped over the ditch vs. Ben jumped towards the ditch）。他們需要選擇其中一句句子用以正確描述練習中的圖片而系統亦會即時向他們提供反饋。SF組（圖1）的反饋包括匹配練習中的圖片的圖像模組及對該模組的文字解釋。。RF組的反饋是提供該介詞在文法上的用法及一般規則並同時給予三個相關例句。CF組則只會獲告知他們在練習中的選擇是正確或錯誤。四個組別都會經歷前期測驗，訓練（各組的訓練時間相約）和後期測驗。測驗由兩個部份組成：句子層級上的填充選擇測驗及翻譯測驗。 / 調查結果顯示，四個組別在前期測驗中的表現並沒有顯著的差異。但在後期測驗中，三個有接受介詞訓練的組別和對照組之間有著的顯著的差異。這一發現表明了研究中的電腦導師對教導介詞有正面的影響。而不同反饋組別（SF，RF和CF組）在後期測驗中則有著邊緣顯著差異，而這差異僅來自翻譯測驗。這結果顯示當研究參加者接受對比教學及圖像刺激時，提供予他們的信息數量及種類對理解型的測驗並沒有太大影響；相反，對運用型的測驗則非常顯著的效果。，當他們被指示具有最小配對模式，也與畫面刺激的幫助。[附圖] / Wong, Man Ho Ivy. / Thesis M.Phil. Chinese University of Hong Kong 2015. / Includes bibliographical references (leaves 141-160). / Abstracts also in Chinese. / Title from PDF title page (viewed on 18, October, 2016). / Wong, Man Ho Ivy. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only.

Yin, Ling = 銅基化合物薄膜太陽能電池 : 新透明導電電極與缺陷物理 / 尹苓. / Thesis Ph.D. Chinese University of Hong Kong 2015. / Includes bibliographical references. / Abstracts also in Chinese. / Title from PDF title page (viewed on 26, October, 2016). / Yin, Ling = Tong ji hua he wu bo mo tai yang neng dian chi : xin tou ming dao dian dian ji yu que xian wu li / Yin Ling.

I examine the firms with different hedging needs in their investment and financing strategy during the financial crisis in 2008. I define the hedging needs using the correlation between their cash flow and the industry Q. The firms with positive correlation between the cash flow and industry Q are defined as low hedging needs firms and the firms with negative correlation between the cash flow and industry Q are defined as high hedging needs firms. The low hedging needs firms have similar investment growth as high hedging needs firms before the crisis but significantly less investment growth during the crisis. However, the impact of financial crisis on firms with low hedging needs is smaller. Those firms efficiently respond to the decline of investment opportunity during crisis period by cutting capital expenditures since the capital expenditure in crisis are associated with lower profitability. And the empirical results support that there is a mix of supply shock and demand shock during financial crisis in 2008. / Tao, Xiaojue. / Thesis M.Phil. Chinese University of Hong Kong 2015. / Includes bibliographical references (leaves 22-24). / Title from PDF title page (viewed on 11, October, 2016). / Tao, Xiaojue.

As a growing developing economy, China impressed the world by its astonishing growth rate and fast structural transformation. Mean-while, the openness also increased a lot after China adopted the reform and opening up policy in 1978. This paper first constructs real GDP, capital and labor series in China from 1952 to 2012 following instructions by some scholars, and then uses the simple growth accounting method to show that the agricultural productivity growth plays an important role in structural transformation: higher agricultural productivity growth accompanies with further structural transformation, or, lower agricultural employment share. And the role does not change much when the openness of China becomes higher, inconsistent from the outcome of Matsuyama (1992) at a first glance. / 作為一個在成長的發展中國家，中國驚人的增長率和快速的結構轉型給世界留下了深刻的印象。同時，在中國1978 年實行改革開放的政策後，中國的開放程度也大大增加。本文根據一些學者提供的思路構建了中國1952 至2012 年的實際GDP、勞動力及資本的時間序列，然後用簡單增長核算方法展示了農業的生產率增長在結構轉型方面的重要角色：更高的農業相對生產率增長會伴隨更深的結構轉型。當中國的開放程度變高時，這種角色沒有太多改變。初步地看，這和Matsuyama (1992)文中的描述並不一致。 / Chen, Xilu. / Thesis M.Phil. Chinese University of Hong Kong 2015. / Includes bibliographical references (leaves 83-87). / Abstracts also in Chinese. / Title from PDF title page (viewed on 11, October, 2016). / Detailed summary in vernacular field only.