This book delivers the state of the art in deep learning (DL) methods hybridized with evolutionary computation (EC). Over the last decade, DL has dramatically reformed many domains: computer vision, speech recognition, healthcare, and automatic game playing, to mention only a few. All DL models, using different architectures and algorithms, utilize multiple processing layers for extracting a hierarchy of abstractions of data. Their remarkable successes notwithstanding, these powerful models are facing many challenges, and this book presents the collaborative efforts by researchers in EC to solve some of the problems in DL. EC comprises optimization techniques that are useful when problems are complex or poorly understood, or insufficient information about the problem domain is available. This family of algorithms has proven effective in solving problems with challenging characteristics such as non-convexity, non-linearity, noise, and irregularity, which dampen the performance of most classic optimization schemes. Furthermore, EC has been extensively and successfully applied in artificial neural network (ANN) research -from parameter estimation to structure optimization. Consequently, EC researchers are enthusiastic about applying their arsenal for the design and optimization of deep neural networks (DNN). This book brings together the recent progress in DL research where the focus is particularly on three sub-domains that integrate EC with DL: (1) EC for hyper-parameter optimization in DNN; (2) EC for DNN architecture design; and (3) Deep neuroevolution. The book also presents interesting applications of DL with EC in real-world problems, e.g., malware classification and object detection. Additionally, it covers recent applications of EC in DL, e.g. generative adversarial networks (GAN) training and adversarial attacks. The book aims to prompt and facilitate the research in DL with EC both in theory and in practice.
Autonomous robots have been a vision of robotics, artificial intelligence, and cognitive sciences. An important step towards this goal is to create robots that can learn to accomplish amultitude of different tasks triggered by environmental context and higher-level instruction. Early approaches to this goal during the heydays of artificial intelligence research in the late 1980s showed that handcrafted approaches do not suffice and that machine learning is needed. However, off the shelf learning techniques often do not scale into real-time or to the high-dimensional domains of manipulator and humanoid robotics. In this book, we investigate the foundations for a general approach to motor skilllearning that employs domain-specific machine learning methods. A theoretically well-founded general approach to representing the required control structures for task representation and executionis presented along with novel learning algorithms that can be applied in this setting. The resulting framework is shown to work well both in simulation and on real robots.
Keywords of Systemic Vision is meant to be an encyclopedia of Systemic and Complex science. The word "encyclopedia" refers to definitely acquired information and deep knowledge in a given domain, so that also people who do not belong to the related scientific community can rely on accurate representations of that object. However, Complexity science refers to phenomena, for which it is impossible to find objective information, for reasons such as chaotic evolution, high number of variables, observers' bias etc. Consequently, the expectations of an encyclopedia of Complexity science should be different from those of a traditional one: this work highlights the incertitude that has always affected complex phenomena, as well as the ways of dealing with this incertitude in very different scientific domains.
This book proposes a comprehensible, context-sensitive and flexible framework for the development of pedagogy for autonomy in language education. The "framework" metaphor highlights the effort to identify structuring elements in the authors' stance towards pedagogy for autonomy, which fall into three domains -the context, the learner, and the teacher. In each domain, the authors raise ethical, conceptual and practical issues that are crucial to their perspective and offer a basis for reflection on learner and teacher development towards learner and teacher autonomy. The book proposes a common definition for learner and teacher autonomy within a vision of education as transformation and empowerment. Pedagogy for autonomy is operationalized through a set of ten general principles.
This book is dedicated to the consolidation and to the expansion of theoretic systems thinking as a necessary integration of the general reductionist and analytical attitude dominant in our culture. Reductionism and analytical approaches have produced significant results in many fields of contemporary knowledge giving a great contribution to relevant scientific discoveries and to their technological application, but their validity has been improperly universalized as the only and best methods of knowledge in every domain. It is nowadays clear that analytical or mereological approaches are inadequate to solve many problems and that we should introduce - or support the diffusion of - new concepts and different research attitudes. A good candidate to support such a shift is the well known theoretical approach based on the concept of "system" that no more considers the elementary constituents of an object, but the entity emerging from the relations and interactions among its elementary parts. It becomes possible to reconstruct several domains, both philosophical and scientific, from the systemic point of view, introducing fresh ideas in the research in view of a general rational vision of the world on more comprehensive basis. This book contributes to the diffusion and evolution of systemic thinking by focusing on two main objectives: developing and updating the systemic approach in disciplines currently using it and introducing the systemic perspective in humanistic disciplines, where the approach is not widely used. The Systemic Turn in Human and Natural Sciences: A Rock in the Pond is comprised of ten chapters. The chapter authors adopt a trans-disciplinary perspective, consisting in the recognition and harmonization of the special outlooks that together, within the general systemic paradigm, gives an ideal unity to the book.
A key driving factor for biometrics is the widespread national and international depl- ment of biometric systems that has been initiated in the past two years and is about to accelerate. While nearly all current biometric deployments are government-led and pr- cipally concerned with national security and border control scenarios, it is now apparent that the widespread availability of biometrics in everyday life will also spin out an ev- increasing number of (private) applications in other domains. Crucial to this vision is the management of the user's identity, which does not only imply the creation and update of a biometric template, but requires the development of instruments to properly handle all the data and operations related to the user identity. COST Action 2101 on Biometrics for Identity Documents and Smart Cards has - erated as a valuable and effective platform for close collaboration of European sci- tists from academia and industry researching biometrics for identity documents and smartcards. This has led to the continuous advances achieved in various classes of biometrics and their implementations in the identity management domain. These c- tributions to knowledge in this field were first presented at the First European Wo- shop on Biometrics and Identity Management (BioID 2008) organized in Roskilde, Denmark during May 7-9, 2008.
Autonomous robots have been a vision of robotics, artificial intelligence, and cognitive sciences. An important step towards this goal is to create robots that can learn to accomplish a multitude of different tasks triggered by environmental context and higher-level instruction. Early approaches to this goal during the heydays of artificial intelligence research in the late 1980s showed that handcrafted approaches do not suffice and that machine learning is needed. However, off the shelf learning techniques often do not scale into real-time or to the high-dimensional domains of manipulator and humanoid robotics. In this book, we investigate the foundations for a general approach to motor skill learning that employs domain-specific machine learning methods. A theoretically well-founded general approach to representing the required control structures for task representation and execution is presented along with novel learning algorithms that can be applied in this setting. The resulting framework is shown to work well both in simulation and on real robots.
An apt description of this book appears on the back cover of the book. This descrition was written by Professor Marié-Heleen Coetzee (DTech) and we quote:'Arthur Lessac's passing in 2011 has not only left the Lessac community grieving for a remarkable man, but has brought into sharp focus the need for a continuation of his vision. This book does exactly that. Beyond furthering his vision, this book brings together artist-scholars who extend his vision into multi-disciplinary domains. By weaving together diverging theoretical and contextual terrains, the book stands testimony to the richness and malleability of Lessac Kinesensics.The book organically develops from a layered exposition of the work to applications of the work in the domains of performance and well-being. The central epistemological principle underpinning the book is in viewing the bodymindvoice as a mode of 'knowing' and 'being-in-the-world'; simultaneously being-thinking in and through the work in accordance with the parameters of a specific learning experience or research project. The book offers an array of practice-based methodologies to compliment this principle, contributing not only to scholarship on Lessac Kinsensics, but also to the broader domain of research methodologies in/for the arts. In weaving together theoretical terrains and autobiographical narratives, the book offers a reading experience that is simultaneously deeply (and, at times, movingly) immersive and critically reflexive. In terms of praxis, this book foregrounds the ways in which Lessac Kinesensics fosters self-use, self-teaching and self-care - offering ways in which we can better interrogate our inner worlds, extend our understanding of others and optimize our ways of engaging with the external world.The book is a manifesto of hope: using Lessac Kinesensics to activiate and collectively vision an artistic and pedagogical future greated towards human flourishing.'
Interoperability of enterprises is one of the main requirements for economical and industrial collaborative networks. Enterprise interoperability (EI) is based on the three domains: architectures and platforms, ontologies and enterprise modeling. This book presents the EI vision of the 'Grand Sud-Ouest' pole (PGSO) of the European International Virtual Laboratory for Enterprise Interoperability (INTEROP-VLab). It includes the limitations, concerns and approaches of EI, as well as a proposed framework which aims to define and delimit the concept of an EI domain. The authors present the basic concepts and principles of decisional interoperability as well as concept and techniques for interoperability measurement. The use of these previous concepts in a healthcare ecosystem and in an extended administration is also presented.