Abstract : In recent years, Information and Communication Science and Technology have witnessed awesome advances thanks to the groundbreaking nature of new materials, calculation processes and data sources. 'Grey box' computers now only represent a small proportion of calculation resources and data sources. Indeed, more than 80% of processors are today integrated into various sophisticated devices. The number of sensors integrated into components with processing and signal transmission units has significantly increased. Each sensor is an active node in a system whose local processing capabilities make it possible to aggregate, sort and filter data or carry out more sophisticated processing. Human-computer interaction has also significantly evolved. It is no longer simply confined to the traditional 'screen, keyboard, mouse' triplet but permeates our everyday objects and activities. User Interfaces (UIs) are no more limited to graphics, neither to static contexts of use. Rather, they become multimodal and capable of adaptation to dynamic contexts of use. They migrate from one interactive space to another as long as the user moves. Whilst this vision is exciting from a usage point of view, it raises grand challenges for the engineering of such UIs. These challenges are as great as calculatory devices are powerful and therefore allow merging and mining huge databases. Anticipating and overcoming the risk of system hijacks are also part of these issues. Transparency becomes a highly valuable quality for ensuring better access to resources at all levels of the system or corporation ('virtual', 'in-network' etc.) making resources vulnerable to threats and attacks. Due to the wide range of risks, from economic intelligence to protecting personal data, it is necessary to find the right balance between ensuring data protection and transparency of access to new autonomous resources in open environments. New challenges have emerged such as those related to managing data access, ethics and the well known 'precautionary principle', 'Big data', 'Big Brother'; the list is endless. With such challenges, the growing diversity of dynamic services and of smart objects raises new issues in the design, development and execution of software applications. These applications must be able to adapt to a software and hardware infrastructure that continuously and unpredictably changes. Prime examples are applications that follow the user as he/ she moves, such as those used in mobile phones, cars, houses, etc. which provide him/ her with permanent access to services over a prolonged period of time. As such, the application must face to variations in the context of use, and nevertheless ensure its quality of service. Having simplified the distribution of software applications, middleware have now to facilitate the design of these applications by providing them with the ability to adapt. They must provide software mechanisms at runtime which guarantee the permanent adaptation of the application to a changing context of use. These challenges will therefore increase when faced with new usage in increasingly diverse, variable and unpredictable contexts of use. The present book 'Computer science and ambient intelligence: from sensors to applications' is an outcome of two CNRS 'ambient intelligence' schools organized in July of 2009 and 2011. In line with the ethos of this school, the present work aims to inform the lay reader of the challenges posed by this new field of research. Taking a holistic view, it covers various levels of abstraction, ranging from fundamental to advanced concepts and bringing together the contributions of various specialists in the field, the majority of whom have carried out their research within the school. This book features the main areas of computer science concerned with ambient intelligence (e.g. human-computer interaction, middleware, networks and information systems, etc.). It is a multidisciplinary advance with contributions coming from intelligent materials and ethics, the aim of which is to demonstrate the importance of integrated research, based on social sciences and technological advances. Such research is multi-disciplinary with the aim to mobilize and bring together expertise from each field to develop new theories. This book also pays tribute to the field's wide spectrum of applications with chapters focusing on health, transport and even tourism. The aim of the current work is not pedagogical; rather it is designed to provide a stimulating perspective to attest to the challenge of teaching within current frameworks due to the paradigm's interdisciplinary and contemporary nature and the lack of structures, platforms and generic materials. Recent initiatives such as 'FabLabs' are surely part of a response to this. We would like to warmly thank all the authors who have contributed to the creation of this book. We also sincerely hope that you have as much enjoyment reading their contributions as we have had in listening to their talks during both editions of the Ambient Intelligent school.