A Survey of Image Gradient Inversion Against Federated Learning

In order to preserve data privacy while fully utilizing data from different owners, federated learning is believed to be a promising approach in recent years. However, aiming at federated learning in the image domain, gradient inversion techniques can reconstruct the input images on pixel-level only by leaked gradients, without accessing the raw data, which makes federated learning vulnerable to the attacks. In this paper, we review the latest advances of image gradient inversion techniques and evaluate the impact of them to federated learning from the attack perspective. We use eight models and four datasets to evaluate the current gradient inversion techniques, comparing the attack performance as well as the time consumption. Furthermore, we shed light on some important and interesting directions of gradient inversion against federated learning.<br>

Download Full-text

A Survey of Image Gradient Inversion Against Federated Learning

10.36227/techrxiv.18254723.v1 ◽

2022 ◽

Author(s):

Zhaohua Li ◽

Le Wang ◽

Guangyao Chen ◽

Muhammad Shafq ◽

zhaoquan Gu

Keyword(s):

Data Privacy ◽

Raw Data ◽

Image Domain ◽

Time Consumption ◽

Image Gradient ◽

Inversion Techniques ◽

The Impact ◽

Shed Light

Download Full-text

It Is More Than Language

The Oxford Handbook of Deaf Studies in Learning and Cognition ◽

10.1093/oxfordhb/9780190054045.013.1 ◽

2020 ◽

pp. 79-98

Author(s):

Dianne Toe ◽

Louise Paatsch ◽

Amy Szarkowski

Keyword(s):

Specific Area ◽

Cooperative Interaction ◽

Social And Emotional Development ◽

Children And Young People ◽

Social And Emotional ◽

The Social ◽

Pragmatic Skills ◽

Social Use ◽

The Impact ◽

Shed Light

Deaf and hard-of-hearing (DHH) children who use spoken language face unique challenges when communicating with others who have typical hearing, particularly their peers. In such contexts, the social use of language has been recognized as an area of vulnerability among individuals in this population and has become a focus for research and intervention. The development of pragmatic skills intersects with many aspects of child development, including emotional intelligence and executive function, as well as social and emotional development. While all these areas are important, they are beyond the scope of this chapter, which highlights the impact of pragmatics on the specific area of cognition. Cognitive pragmatics is broadly defined as the study of the mental processes involved in the understanding of meaning in the context of a cooperative interaction. This chapter explores how DHH children and young people construe meaning in the context of conversations and expository interactions with their peers. The chapter aims to examine the role played by the cognitive processes of making inferences and comprehending implicature, within the overall display of pragmatic skills. Further, the authors use this lens in the analysis of interactions between DHH children and their peers in order to shed light on the development of pragmatic skills in children who are DHH.

Download Full-text

The effect of Medicare Part D on prescription drug composition and demand

Journal of Economic Studies ◽

10.1108/jes-08-2013-0109 ◽

2015 ◽

Vol 42 (2) ◽

pp. 170-185 ◽

Cited By ~ 4

Author(s):

David Zimmer

Keyword(s):

Prescription Drug ◽

Medicare Part D ◽

Medicare Part ◽

Part D ◽

Content Type ◽

The Us ◽

Upper Respiratory ◽

Nationally Representative ◽

The Impact ◽

Shed Light

Purpose – The US Medicare Modernization Act of 2003 introduced optional prescription drug coverage, beginning in 2006, widely known as Medicare Part D. This paper uses up-to-date nationally representative survey data to investigate the impact of Part D not only on drug spending and consumption, but also on the composition of drug consumption. The paper aims to discuss these issues. Design/methodology/approach – Specifically, the paper investigates whether Part D impacted the number of therapeutic classes for which drugs were prescribed, and also whether Part D lead to increased usage of drugs for specific medical conditions that typically receive drug-intensive therapies. Findings – In addition to confirming findings from previous studies, this paper shows that Part D increased the number of therapeutic classes to which seniors receive drugs by approximately four classes. Part D also lead to increased usage of drugs used to treat upper respiratory disease, hypertension, and diabetes. Originality/value – While mostly concurring with previous studies on the spending impacts of Part D, this paper is the first to shed light on other impacts of Part D, specifically with respect to its impact on therapeutic classes for which drugs are prescribed.

Download Full-text

Publishing Productivity in Sport Psychology 1970–2000: An Exploratory Examination of the Lotka-Price Law

Journal of Sport and Exercise Psychology ◽

10.1123/jsep.25.4.477 ◽

2003 ◽

Vol 25 (4) ◽

pp. 477-483 ◽

Cited By ~ 10

Author(s):

Joseph Baker ◽

Jennifer Robertson-Wilson ◽

Whitney Sedgwick

Keyword(s):

Sport Psychology ◽

Scientific Productivity ◽

Eminent Scientist ◽

Research Papers ◽

Limited Support ◽

The Impact ◽

Published Research ◽

Shed Light

The current study examined whether the distribution of published research papers in the field of sport psychology followed the Lotka-Price Law of scientific productivity. All authors who had published articles in five sport psychology journals from 1970 to 2000 were considered. The impact of those authors was determined by the total number of published papers in all journals. Results provided limited support for the Lotka-Price Law; however, it appeared that the field of sport psychology was less elitist than other fields. Although these findings suggest that productivity in this field is similar to that in other fields of science, more research is needed to shed light on the role of the eminent scientist and the average researcher in the advancement of knowledge in sport psychology.

Download Full-text

Benznidazole Treatment following Acute Trypanosoma cruzi Infection Triggers CD8+ T-Cell Expansion and Promotes Resistance to Reinfection

Antimicrobial Agents and Chemotherapy ◽

10.1128/aac.46.12.3790-3796.2002 ◽

2002 ◽

Vol 46 (12) ◽

pp. 3790-3796 ◽

Cited By ~ 25

Author(s):

Bianca Perdigão Olivieri ◽

Vinícius Cotta-de-Almeida ◽

Tania Araújo-Jorge

Keyword(s):

T Lymphocytes ◽

Trypanosoma Cruzi ◽

Immune Regulation ◽

Immune Dysregulation ◽

Direct Role ◽

Parasite Replication ◽

The Impact ◽

Shed Light ◽

Cruzi Infection

ABSTRACT Many studies have shed light on the mechanisms underlying both immunoprotection and immune dysregulation arising after Trypanosoma cruzi infection. However, little is known about the impact of benznidazole (N-benzyl-2-nitroimidazole acetamide), the drug available for clinical treatment of the infection, on the immune system in the infected host. In the present study we investigated the effect of benznidazole therapy on the lymphoid compartment during the course of experimental T. cruzi infection. Although amelioration of a variety of clinical and parasitological signs was observed in treated mice, amelioration of splenocyte expansion was not detected. Interestingly, this sustained splenomegaly observed in benznidazole-treated mice showed a preferential expansion of CD8+ T lymphocytes. Moreover, although benznidazole treatment blocked the expansion of recently activated CD4+ and CD8+ T cells seen in infected hosts, benznidazole treatment led to a selective expansion of effector and memory CD8+ T lymphocytes in association with a lower rate of apoptosis. In addition, the surviving treated animals were protected from reinfection. Together, these data suggest that, in addition to its well-known direct role in blocking parasite replication in vivo, benznidazole appears to directly affect immune regulation in T. cruzi-infected hosts.

Download Full-text

Evaluating E-Book Effectiveness and the Impact on Student Engagement

Journal of Marketing Education ◽

10.1177/02734753211035162 ◽

2021 ◽

pp. 027347532110351

Author(s):

Adam C. Merkle ◽

Linda K. Ferrell ◽

O. C. Ferrell ◽

Joe F. Hair

Keyword(s):

Academic Performance ◽

Student Engagement ◽

Student Perceptions ◽

Perceived Effectiveness ◽

Spring Semester ◽

Digital Disruption ◽

The Impact ◽

The Relationship ◽

Shed Light

Marketing curricula are experiencing a digital disruption as e-books and other electronic educational resources replace print textbooks. This study investigates student perceptions about the effectiveness of print textbooks and e-books. Specifically, we focus on the perceived effectiveness of e-books and the impact on student engagement. A field-based quasi-experiment was conducted with a sample of 259 students in the Fall semester, and a follow-up sample of 395 students in the Spring semester. The results show a diverse impact of e-books on student engagement. Some aspects of engagement are positively affected while other aspects of student engagement exhibit a neutral or negative leaning impact. The findings also reflect significant variation in e-book effectiveness depending on the course. Finally, we find that e-books moderate the relationship between textbook effectiveness and academic performance engagement. Highly effective e-books result in higher levels of academic performance engagement. Collectively these findings shed light on the current situation and provide a foundation for additional research to further our understanding about e-book effectiveness and its relationship to student engagement.

Download Full-text

Solutions to Big Data Privacy and Security Challenges Associated With COVID-19 Surveillance Systems

Frontiers in Big Data ◽

10.3389/fdata.2021.645204 ◽

2021 ◽

Vol 4 ◽

Author(s):

Vibhushinie Bentotahewa ◽

Chaminda Hewage ◽

Jason Williams

Keyword(s):

Big Data ◽

Data Privacy ◽

Personal Data ◽

Surveillance Systems ◽

Privacy Rights ◽

Personal Privacy ◽

Privacy And Security ◽

Digital Surveillance ◽

Feasible Solutions ◽

The Impact

The growing dependency on digital technologies is becoming a way of life, and at the same time, the collection of data using them for surveillance operations has raised concerns. Notably, some countries use digital surveillance technologies for tracking and monitoring individuals and populations to prevent the transmission of the new coronavirus. The technology has the capacity to contribute towards tackling the pandemic effectively, but the success also comes at the expense of privacy rights. The crucial point to make is regardless of who uses and which mechanism, in one way another will infringe personal privacy. Therefore, when considering the use of technologies to combat the pandemic, the focus should also be on the impact of facial recognition cameras, police surveillance drones, and other digital surveillance devices on the privacy rights of those under surveillance. The GDPR was established to ensure that information could be shared without causing any infringement on personal data and businesses; therefore, in generating Big Data, it is important to ensure that the information is securely collected, processed, transmitted, stored, and accessed in accordance with established rules. This paper focuses on Big Data challenges associated with surveillance methods used within the COVID-19 parameters. The aim of this research is to propose practical solutions to Big Data challenges associated with COVID-19 pandemic surveillance approaches. To that end, the researcher will identify the surveillance measures being used by countries in different regions, the sensitivity of generated data, and the issues associated with the collection of large volumes of data and finally propose feasible solutions to protect the privacy rights of the people, during the post-COVID-19 era.

Download Full-text

The Role of Education in Human Development in Iraq for the Period 1980-2012

Journal of University of Human Development ◽

10.21928/juhd.v2n2y2016.pp251-276 ◽

2016 ◽

Vol 2 (2) ◽

pp. 251

Author(s):

Sabah Faihan Mahmood ◽

Yassen Taha Mahmood

Keyword(s):

Human Development ◽

Human Development Index ◽

Economic Activities ◽

Development Index ◽

Level Of Knowledge ◽

Index Value ◽

The Impact ◽

The Relationship ◽

Shed Light

Human Development aims to enlarge choices in front of people by improving the level of health, education, and income; this means that this process will upgrade both the economic and social development.In other words, human development aimes to raise the average of age and this requires the advancement of the health aspect, raise the level of knowledge and this requires the advancement of the educational aspect of all kinds., and raise the standard of living, and this requires the advancement of the economic aspect by providing the necessary jobs and promote economic activities. The study focus on the relationship between education and human development which has great importance as a mean to determine the impact of education on human development. The research seeks to achieve a set of objectives, including: Review the concept of human development and its basic elements, shed light on the reality of development in Iraq and follow the path of its development, and find out the role of education in influencing human development through the changes taking place in it and its impact on increase or decrease human development index during the period of the research. The research found set of results, the most important were the important effect of the education index on the level of human development index represented by HDI. Iraq had a good educational system in the eighties and nineties, reaching good education index value for the year (1990) which was (0.890), making the human development index in Iraq in the highest level and the value of the Human Development Index (0.759) in the first report issued by the United Nations in the year (1990). when the education index fall back there was negatively impact on the value of human development index in Iraq Directory, so when the education index value became (0.721) , the value of the Human Development Index was (0.590) in the year 2011 . This means that the value of the human development index decrease in recent years, although of improvement in the level of health, and the average per capita GDP in Iraq, and this illustrates the significant role of education on the human development process.

Download Full-text

Study of a Context Quality Model for UbiPri Middleware

KnE Engineering ◽

10.18502/keg.v5i6.7075 ◽

2020 ◽

Author(s):

Diandre de Paula ◽

Daniel Saraiva ◽

Romeiro Natália ◽

Nuno Garcia ◽

Valderi Leithardt

Keyword(s):

Quality Indicators ◽

Data Privacy ◽

Context Information ◽

Adaptation Process ◽

Context Aware ◽

Quality Model ◽

Context Aware Computing ◽

Quality Of Context ◽

The Impact

With the growth of ubiquitous computing, context-aware computing-based applications are increasingly emerging, and these applications demonstrate the impact that context has on the adaptation process. From the context, it will be possible to adapt the application according to the requirements and needs of its users. Therefore, the quality of the context information must be guaranteed so that the application does not have an incorrect or unexpected adaptation process. But like any given data, there is the possibility of inaccuracy and/or uncertainty and so Quality of Context (QoC) plays a key role in ensuring the quality of context information and optimizing the adaptation process. To guarantee the Quality of Context it is necessary to study a quality model to be created, which will have the important function of evaluating the context information. Thus, it is necessary to ensure that the parameters and quality indicators to be used and evaluated are the most appropriate for a given type of application. This paper aims to study a context quality model for the UbiPri middleware, defining its quality indicators to ensure its proper functioning in the process of adaptation in granting access to ubiquitous environments. Keywords: QoC, Model, Context-Aware, Data, Privacy

Download Full-text

CDC: Classification Driven Compression for Bandwidth Efficient Edge-Cloud Collaborative Deep Learning

Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence ◽

10.24963/ijcai.2020/467 ◽

2020 ◽

Author(s):

Yuanrui Dong ◽

Peng Zhao ◽

Hanqiao Yu ◽

Cong Zhao ◽

Shusen Yang

Keyword(s):

Deep Learning ◽

Classification Accuracy ◽

Data Privacy ◽

Privacy Preservation ◽

Fine Tuning ◽

Quantization Scheme ◽

Raw Data ◽

Resource Limited ◽

Lower Accuracy ◽

Bandwidth Consumption

The emerging edge-cloud collaborative Deep Learning (DL) paradigm aims at improving the performance of practical DL implementations in terms of cloud bandwidth consumption, response latency, and data privacy preservation. Focusing on bandwidth efficient edge-cloud collaborative training of DNN-based classifiers, we present CDC, a Classification Driven Compression framework that reduces bandwidth consumption while preserving classification accuracy of edge-cloud collaborative DL. Specifically, to reduce bandwidth consumption, for resource-limited edge servers, we develop a lightweight autoencoder with a classification guidance for compression with classification driven feature preservation, which allows edges to only upload the latent code of raw data for accurate global training on the Cloud. Additionally, we design an adjustable quantization scheme adaptively pursuing the tradeoff between bandwidth consumption and classification accuracy under different network conditions, where only fine-tuning is required for rapid compression ratio adjustment. Results of extensive experiments demonstrate that, compared with DNN training with raw data, CDC consumes 14.9× less bandwidth with an accuracy loss no more than 1.06%, and compared with DNN training with data compressed by AE without guidance, CDC introduces at least 100% lower accuracy loss.

Download Full-text