Search

Federated learning (FL) is an emerging machine learning technique where machine learning models are trained in a decentralized manner. The main advantage of this approach is the data privacy it provides because the data are not processed in a centralized device. Moreover, the local client models are aggregated on a server, resulting in a global model that has accumulated knowledge from all the different clients. This approach, however, is vulnerable to attacks because clients can be malicious or malicious actors may interfere within the network. In the first case, these types of attacks may refer to data or model poisoning attacks where the data or model parameters, respectively, may be altered. In this paper, we investigate the data poisoning attacks and, more specifically, the label-flipping case within a federated learning system. For an image classification task, we introduce two variants of data poisoning attacks, namely model degradation and targeted label attacks. These attacks are based on synthetic images generated by a generative adversarial network (GAN). This network is trained jointly by the malicious clients using a concatenated malicious dataset. Due to dataset sample limitations, the architecture and learning procedure of the GAN are adjusted accordingly. Through the experiments, we demonstrate that these types of attacks are effective in achieving their task and managing to fool common federated defenses (stealth). We also propose a mechanism to mitigate these attacks based on clean label training on the server side. In more detail, we see that the model degradation attack causes an accuracy degradation of up to 25%, while common defenses can only alleviate this for a percentage of ∼5%. Similarly, the targeted label attack results in a misclassification of 56% compared to 2.5% when no attack takes place. Moreover, our proposed defense mechanism is able to mitigate these attacks. © 2023 by the authors.

Meeting, socializing and conversing online with a group of people using teleconferencing systems is still quite different from the experience of meeting face to face. We are abruptly aware that we are online and that the people we are engaging with are not in close proximity. Analogous to how talking on the telephone does not replicate the experience of talking in person. Several causes for these differences have been identied and we propose inspiring and innovative solutions to these hurdles in attempt to provide a more realistic, believable and engaging online conversational experience. We present the distributed and scalable framework REVERIE that provides a balanced mix of these solutions. Applications build on top of the REVERIE framework will be able to provide interactive, immersive, photo-realistic experiences to a multitude of users that for them will feel much more similar to having face to face meetings than the experience offered by conventional teleconferencing systems.

Meeting, socializing and conversing online with a group of people using teleconferencing systems is still quite differ- ent from the experience of meeting face to face. We are abruptly aware that we are online and that the people we are engaging with are not in close proximity. Analogous to how talking on the telephone does not replicate the experi- ence of talking in person. Several causes for these differences have been identified and we propose inspiring and innova- tive solutions to these hurdles in attempt to provide a more realistic, believable and engaging online conversational expe- rience. We present the distributed and scalable framework REVERIE that provides a balanced mix of these solutions. Applications build on top of the REVERIE framework will be able to provide interactive, immersive, photo-realistic ex- periences to a multitude of users that for them will feel much more similar to having face to face meetings than the expe- rience offered by conventional teleconferencing systems.

Irrespective of whether we use economic or societal metrics, the Internet is one of the most important technical infrastructures in existence today. It will serve as a catalyst for much of our innovation and prosperity in the future. A competitive Europe will require Internet connectivity and services beyond the capabilities offered by current technologies. Future Internet research is therefore a must. The Future Internet Assembly (FIA) is a successful and unique bi-annual conference that brings together participants of over 150 projects from several distinct but interrelated areas in the EU Framework Programme 7. The 20 full papers included in this volume were selected from 40 submissions, and are preceded by a vision paper describing the FIA Roadmap. The papers have been organized into topical sections on the foundations of Future Internet, the applications of Future Internet, Smart Cities, and Future Internet infrastructures.

Irrespective of whether we use economic or societal metrics, the Internet is one of the most important technical infrastructures in existence today. It will be a catalyst for much of our innovation and prosperity in the future. A competitive Europe will require Internet connectivity and services beyond the capabilities offered by current technologies. Future Internet research is therefore a must. This book is published in full compliance with the Open Access publishing initiative; it is based on the research carried out within the Future Internet Assembly (FIA). It contains a sample of representative results from the recent FIA meetings spanning a broad range of topics, all being of crucial importance for the future Internet. The book includes 32 contributions and has been structured into the following sections, each of which is preceded by a short introduction: Foundations: architectural issues; socio-economic issues; security and trust; and experiments and experimental design. Future Internet Areas: networks, services, and content; and applications.

The Internet is a remarkable catalyst for creativity, collaboration and innovation, providing us today with amazing possibilities that just two decades ago would have been impossible to imagine. Our challenge today is to prepare a trip into the future: what will be the Internet in ten or twenty years from now and what more amazing things will it offer to people? In order to see what the future will bring, we first need to consider some important challenges that the Internet faces today. European scientists proved that they are at the forefront of Internet research already since the invention of the web. But the challenges are huge and complex and cannot be dealt with in isolation. The European Future Internet Assembly is the vehicle to a fruitful scientific dialogue, bringing together the different scientific disciplines that contribute to the Future Internet development. Until now, scientists from more than 90 research projects were funded with around 300 million euros under the 7th Framework Programme. Another 400 million euros will be made available in the near future. These amounts coupled with private investments bring the total investment to more than a billion euros, showing Europe’s commitment to address the challenges of the future Internet. This book is a peer-reviewed collection of scientific papers addressing some of the challenges ahead that will shape the Internet of the Future. The selected papers are representative of the research carried out by EU-funded projects in the field. European scientists are working hard to make the journey to the Future Internet as exciting and as fruitful as was the trip that brought us the amazing achievements of today. We invite you to read their visions and join them in their effort so Europe can fully benefit from the exciting opportunities in front of us.

Socio-economics aims to understand the interplay between the society, economy, markets, institutions, self-interest, and moral commitments. It is a multidisciplinary field using methods from economics, psychology, sociology, history, and even anthropology. Socio-economics of networks have been studied for over 30 years, but mostly in the context of social networks instead of the underlying communication networks. The aim of this paper is to present and discuss challenges and perspectives related to “socio-economic” issues in the Future Internet. It is hoped that this will lead to new insights on how to structure the architecture and services in the Internet of the future.

Socio-economic aspects play an increasingly important role in the Future Internet. To enable a TripleWin situation for the involved players, i.e. the end users, the ISPs and telecommunication operators, and the service providers, a new, incentive-based concept is proposed referred to as Economic Traffic Management (ETM). It aims at reducing costs within the network while improving the Quality-of-Experience (QoE) for end users. In particular, peer-to-peer (P2P) overlay applications generate a large amount of costs due to inter-domain traffic. ETM solution approaches have to take into account (a) the traffic patterns stemming from the overlay application, (b) the charging models for transit traffic, and (c) the applicability and efficiency of the proposed solution. The complex interaction between these three components and its consequences is demonstrated on selected examples. As a result it is shown that different ETM approaches have to be combined for an overall solution. To this end, the paper derives functional and non-functional requirements for designing ETM and provides a suitable architecture enabling the implementation of a TripleWin solution.

The concept of Economic Traffic Management (ETM) encompasses various techniques for optimizing overlay networks considering both, underlay and overlay networks’ performance requirements as well as the resulting economic implications for ISPs. This work presents several mechanisms through an overall ETM System (ETMS), identifying the possibility for synergies between mechanisms, both in the sense of complementarity of decision grounds and in the sense of functionality and components employed thereby. The paper describes the core ETMS architecture and how various mechanisms are instantiated. It continues with the discussion of the flexibility and modularity of this architecture, allowing for the accommodation of synergies. Finally, it presents selected results from the test-bed trials of the ETMS and a dedicated discussion on incentives behind these ETM mechanisms.