4,103 research outputs found
ko-ax photo [Selected by Tate Curator of Photography, Simon Baker]
No full textko-ax photo was an open submission competition selected by Simon Baker (Curator of Photography, Tate), Sue Steward (Photography critic, Telegraph, Observer, Guardian, BBC) and John Gill (Curator Brighton Photo Biennial and Founder Photoworks). The 10 artists selected all presented fascinating artworks that conceal narratives and ask questions of the viewer. Questions of beauty, family, decay and fantasy were all explored across over 50 works.</p
The people behind the papers – Jason Ko and Daniel Lobo
No full textPlanarians grow when they are fed and shrink during periods of starvation. However, it is unclear how they maintain appropriate body proportions as their size changes. A new paper in Development investigates the differences between growth and shrinkage dynamics and builds a mathematical model to explore the mechanisms underpinning these two processes. To learn more about the story behind the paper, we caught up with first author, Jason Ko, and corresponding author, Daniel Lobo, Associate Professor at the University of Maryland.https://doi.org/10.1242/dev.20298
Organizational Learning and Marketing Capability Development: A Study of Charity Retailing Operation of British Social Enterprises
No full textSocial enterprise is a hybrid form of profit- and social benefit-seeking organization whereby traditional nonprofit organizations pursue both their social mission and business opportunities. To embrace this new strategic direction shift, the nonprofit organizations need to develop new competences that will enable them to respond to the changes in the business model. The article investigates the learning mechanisms through which social enterprises develop a marketing capability to deploy their resources in the marketplace as the drivers of competitive advantage in their commercial practice. We study eight cases of UK-based charity retailers, in order to address the role of knowledge accumulation, articulation and codification process in the evolution of marketing capability development. We identify, amongst other things that the critical process of organizational learning for social enterprise is to transfer the experience into organization specific knowledge under the social aspects of constraints
Superior QLC Retention Enhancement of a Large Memory Window FEFET Through Gate Stack Engineering
No full text3-D NAND flash cells with ferroelectric field-effect transistors (FEFETs) have gained significant attention due to emerging challenges in 3-D NAND technology. Although metal-insulator-ferroelectric-insulator-semiconductor (MIFIS) structures provide large memory windows (MWs), FEFET is still under extensive research and faces critical issues such as data retention, write endurance, disturbance, variability, scalability, gate stack thickness, write voltage, and thermal stability of FE-HfO2. Here, we propose a novel gate stack, metal-insulator-high k insulator-ferroelectric-insulator-semiconductor (MIKFIS) to address these challenges. The MIKFIS FEFET achieves a large MW of 12.2 V and demonstrates highly enhanced quad-level-cell (QLC) data retention at both 24 C-degrees and 85 C-degrees. Moreover, MIKFIS offers additional benefits, including reduced monoclinic-(m-) phase formation, reduced gate stack thickness, decreased equivalent oxide thickness (EOT), and enhanced FE switching. The origin of the superior retention of MIKFIS is investigated using a revised measurement technique developed to accurately extract polarization.
Ko au te whenua, te whenua ko au – I am the land, the land is me: An autoethnographic investigation of a secondary school teacher’s experience seeking to enrich learning in outdoor education for Māori students.
No full textThis thesis is my story as an outdoor educator, as a researcher, and a co-participant reflecting on my own actions and experiences as well as those of my students. In this autoethnography I share my revelations and tensions in my role as an outdoor education teacher seeking to enrich the experiences of Māori students. Māori culture and history have largely been ignored in the outdoor education classrooms and environments of Aotearoa New Zealand. After teaching the subject for ten years I didn’t perceive that I was perpetuating the same invisibility in my own outdoor education course. Over this time a number of questions that had fermented at the back of my mind came to the fore; ‘why are so few Māori students opting to take outdoor education as a senior secondary school subject?’ and ‘how can I make the subject of outdoor education more desirable and appealing to Māori?’ A place-responsive approach incorporates and values traditional ways of learning through the notion of place and the stories attached to them. The cultural context of learning about and through place has the potential to provide learning opportunities that are relevant and meaningful to all learners but particularly Māori. Place-responsive pedagogies allow outdoor educators to create an environment where language, knowledge, culture and values are normal, valid and legitimate – contexts where Māori students can be themselves. Through this research I have found that the implementation of a place responsive approach has had significant implications for Year twelve outdoor education at Mount Maunganui College. The improvement in Māori student achievement and numbers selecting the subject have been affirming.
Ko au te whenua, te whenua ko au – I am the land, the land is m
Machine Learning Strategy for Predicting Process Variability Effect in Ultra-scaled GAA FET and 3D NAND Flash Devices
No full text학위논문 (박사) -- 서울대학교 대학원 : 공과대학 전기·정보공학부, 2020. 8. 신형철.본 논문에서는 다양한 공정 변동 요인에 의한 영향을 초소형 GAA FET 소자 및 3차원 NAND Flash Memory 소자에서 정확하게 예측하기 위한 기계 학습 접근법을 제시하였다. 공정 변동성 요인에 의한 영향은 로직 소자와 메모리 소자에서 여러가지 신뢰성 문제의 원인으로 작용하며 특히, 로직 및 메모리 소자의 수율을 결정하는 마진을 감소시켜 정확한 예측 및 제어가 필수적이다.
기계학습 시스템은 크게 비지도적 학습(=Unsupervised Learning), 지도적 학습(=Supervised Learning), 강화 학습(=Reinforcement Learning)의 3가지 계열로 구분된다. 이 중, 소자 특성을 분석하고 변동성 영향 예측을 목적으로 하는 경우 정해진 입출력(Training data) 값에 근거하여 회귀론적 방법으로 예측 모델을 학습시키는 지도적 학습 계열의 기계학습 시스템이 가장 적합한 방법이다. 지도적 학습 계열의 기계학습 시스템은 다양한 변동성 요소에 대한 다각도의 소자 특성을 예측하여야 하기 때문에 다중 노드(=Multi-Node, MN)를 갖는 복잡한 알고리즘(e.g., Artificial neural networks)기반의 다중 입력-다중 출력(=Multi-Input/Multi-Output, MIMO)을 통해 제시되었다.
기계학습 시스템의 초기 단계로 단일 트렌지스터의 변동성 요인을 선행 분석하였다. 초소형 GAA (Gall-All-Around) VFET (Vertical FET) 디바이스의 프로세스 변동 (PV)을 사용하여 주요 전기 매개 변수의 변동을 예측하는 정확하고 효율적인 기계 학습 (ML) 방식을 제시하였다. 제안 된 기계 학습 접근법은 3D 확률론적 TCAD 시뮬레이션과 비교했을 때 동일한 정확도와 우수한 효율성을 보여준다. 인공 신경 네트워크 기반 (ANN) 기계 학습 알고리즘은 MIMO (Multi-input-Multi-Output) 예측을 매우 효과적으로 수행 할 수 있다.
기계 학습 시스템의 발전된 단계로써, 3D NAND 플래시 메모리의 주요 전기 매개 변수의 변화를 예측하는 가변성 인식 기계 학습 시스템을 제안한다. 우리는 최초로 인공 신경 네트워크 (ANN) 알고리즘 기반 ML 시스템의 예측 영향 요인 효과의 정확성, 효율성 및 일반성을 검증하였다. 따라서 다양한 변동 원인으로 인한 장치의 주요 전기적 특성 변화가 동시에 통합적으로 예측된다. 이 알고리즘은 3D 확률론적 TCAD 시뮬레이션을 벤치마킹하여 1 % 미만의 예측 오류율과 80 % 이상의 계산 비용 절감을 보여줍니다. 또한, 층수가 증가함에 따라 다양한 구조 조건을 갖는 3 차원 낸드 플래시 메모리의 동작 특성을 예측함으로써 알고리즘의 일반성을 확인할 수 있다.This paper presents a Machine Learning (ML) approach for accurately predicting the effects of various process variation sources on ultra-scaled GAA FET devices and 3D NAND Flash Memories. The effects of process variability sources cause various reliability problems in logic and memory devices. In particular, accurate prediction and control is essential by reducing the margin that determine the yield of logic and memory devices.
The machine learning system is largely divided into three classes: Unsupervised Learning, Supervised Learning, and Reinforcement Learning. Among them, a supervised learning series machine learning system, which uses a regression method to train predictive models based on input and output (Training data) values, is the most suitable method for analyzing device characteristics and predicting variability effects. Since the machine learning system of the supervised learning series needs to predict the characteristics of various devices of various variability sources, it is possible to use multiple input-multiple outputs (MIMO) based on complex algorithms (artificial neural networks) with multiple nodes (MN).
In the early stages of the ML system, the variability sources of a single transistor is analyzed. We propose an accurate and efficient machine learning approach which predicts variations in key electrical parameters using process variations (PV) from ultra-scaled gate-all-around (GAA) vertical FET (VFET) devices. The proposed machine learning approach shows the same accuracy and good efficiency when compared to 3D stochastic Technology-CAD (TCAD) simulation. Artificial Neural Network Based (ANN) machine learning algorithm can perform Multi-input-Multi-Output prediction very effectively.
As an advanced stage of the ML system, we propose a variability-aware ML approach that predicts variations in the key electrical parameters of 3D NAND Flash memories. For the first time, we have verified the accuracy, efficiency, and generality of the predictive impact factor effects of ANN algorithm-based ML systems. ANN-based ML algorithms can be very effective in MIMO prediction. Therefore, changes in the key electrical characteristics of the device caused by various sources of variability are simultaneously and integrally predicted. This algorithm benchmarks 3D stochastic TCAD simulation, showing a prediction error rate of less than 1% as well as a calculation cost reduction of over 80%. In addition, the generality of the algorithm is confirmed by predicting the operating characteristics of the 3D NAND Flash memory with various structural conditions as the number of layers increases.Chapter 1. Introduction 1
1.1. Emergence of Ultra-scaled 3D Device 1
1.2. Increasing Difficulty of Interpreting Variability Issues 5
1.3. Need for Accurate Variability Prediction 10
Chapter 2. Machine Learning System 15
2.1. Introduction 15
2.2. Analysis of Variability through TCAD Simulation 17
2.3. Structure of Machine Learning Algorithm 25
2.4. Summary 35
Chapter 3. Prediction of Process Variation Effect for Ultra-scaled GAA Vertical FET Devices 40
3.1. Introduction 40
3.2 Simulation Structure and Methodology 42
3.3. Results and Discussion 45
3.4. Summary 58
Chapter 4. Prediction of Process Variation Effect for 3D NAND Flash Memories 63
4.1. Introduction 63
4.2 Simulation Structure and Methodology 64
4.3. Results and Discussion 74
4.4. Summary 99
Chapter 5. Conclusion 104
Bibliography 106
Abstract in Korean 111Docto
KNOWLEDGE ACCUMULATION IN ASIAN PUBLIC ADMINISTRATION RESEARCH: A CRITICAL REVIEW
No full textGiven the growing controversy over the relevance of Anglo-Saxon style public administration to developing countries and a greater demand for more context-relevant theories of public administration in Asia, we should expect that Asian scholars achieve a certain level of knowledge growth in line with this controversy and demand. On the basis of the review of 8810 articles published in nine major international journals during 1990-2011, the author found that the number of articles on Asian public administration is very small, and there is no strong pattern of growth in this regard. In addition, there are very few studies adopting a comparative approach covering multiple Asian countries. Copyright (c) 2013 John Wiley & Sons, Ltd.N
Book-in-Common Conversation - Lisa Ko
No full textLisa Ko, author of The Leavers, will speak about her book and writing career in a virtual presentation on Tuesday, March 23 at 7 pm. Lisa Ko is the recipient of the 2016 PEN/Bellwether Prize and the 2017 Barnes and Noble Discover Great New Writers Award. The Leavers was also named best book of the year by NPR
Record Memory Window (12.2 V) and Superior QLC Retention (10 year, 85°C) by Gate Stack Engineering in Ferroelectric FET: from “MIFIS” to “MIKFIS”
No full textDifferentially altered social dominance- and cooperative-like behaviors in Shank2- and Shank3-mutant mice
No full textBackground: Recent progress in genomics has contributed to the identification of a large number of autism spectrum disorder (ASD) risk genes, many of which encode synaptic proteins. Our understanding of ASDs has advanced rapidly, partly owing to the development of numerous animal models. Extensive characterizations using a variety of behavioral batteries that analyze social behaviors have shown that a subset of engineered mice that model mutations in genes encoding Shanks, a family of excitatory postsynaptic scaffolding proteins, exhibit autism-like behaviors. Although these behavioral assays have been useful in identifying deficits in simple social behaviors, alterations in complex social behaviors remain largely untested. Methods: Two syndromic ASD mouse models—Shank2 constitutive knockout [KO] mice and Shank3 constitutive KO mice—were examined for alterations in social dominance and social cooperative behaviors using tube tests and automated cooperation tests. Upon naïve and salient behavioral experience, expression levels of c-Fos were analyzed as a proxy for neural activity across diverse brain areas, including the medial prefrontal cortex (mPFC) and a number of subcortical structures. Findings: As previously reported, Shank2 KO mice showed deficits in sociability, with intact social recognition memory, whereas Shank3 KO mice displayed no overt phenotypes. Strikingly, the two Shank KO mouse models exhibited diametrically opposed alterations in social dominance and cooperative behaviors. After a specific social behavioral experience, Shank mutant mice exhibited distinct changes in number of c-Fos+ neurons in the number of cortical and subcortical brain regions. Conclusions: Our results underscore the heterogeneity of social behavioral alterations in different ASD mouse models and highlight the utility of testing complex social behaviors in validating neurodevelopmental and neuropsychiatric disorder models. In addition, neural activities at distinct brain regions are likely collectively involved in eliciting complex social behaviors, which are differentially altered in ASD mouse models. © 2020, The Author(s).1
- …
