In cross-device private federated learning, differentially private follow-the-regularized-leader (DP-FTRL) has emerged as a promising privacy-preserving method. However, existing approaches assume a semi-honest server and have not addressed the challenge of securely removing this assumption. This is due to its statefulness, which becomes particularly problematic in practical settings where clients can drop out or be corrupted. While trusted execution environments (TEEs) might seem like an obvious solution, a straightforward implementation can introduce forking attacks or availability issues due to state management. To address this problem, our paper introduces a novel server extension that acts as a trusted computing base (TCB) to realize maliciously secure DP-FTRL. The TCB is implemented with an ephemeral TEE module on the server side to produce verifiable proofs of server actions. Some clients, upon being selected, participate in auditing these proofs with small additional communication and computational demands. This extension solution reduces the size of the TCB while maintaining the system's scalability and liveness. We provide formal proofs based on interactive differential privacy, demonstrating privacy guarantee in malicious settings. Finally, we experimentally show that our framework adds small constant overhead to clients in several realistic settings.

In future beyond 5G networks, the demand on the network would be more diversified and personalized. The variety of underlying physical network would also be accelerated by utilization small and heterogeneous cells and the co-existence of cellular networks and other types of wireless networks. User mobility and unexpected application service utilization would highly fluctuate the network demand spatio-temporally. However, per-service and static network slicing considered in 5G networks cannot accommodate such requirements. In this paper, we propose a novel network architecture, User-Oriented Network slicing Architecture (UONA). The proposed architecture has the following two major characteristics. One is that we maintain network slices on a per-user basis, not on a per-service basis. Such high-resolution of network slices would satisfy various and personal requirements from users. The other is decoupling the process of generating network slices into two subprocess, providing subslices by subslice providers and constructing end-to-end user network slices from the subslices by network slice brokers.We introduce the overall design of UONA and explain its advantages, as well as the research challenges to realize it. We also demonstrate the effectiveness of per-user configuration of network subslice by presenting the preliminary numerical evaluation results.

This paper focuses on the covering mechanism in Learning Classifier System (LCS) which generates a new classifier (i.e., an if-then rule) when no classifier matches the input. We propose Local Covering, a niche-based rule generation method for XCS, to reduce the sensitivity of a hyperparameter that determines the generalization probability of initial rules. In Local Covering, the system looks up classifiers in the population and finds the closest match from the input. After that, the selected classifier is copied as the covering classifier and its condition is generalized to cover the input. To integrate the Local Covering method with XCS, XCS-LCPCI (XCS with Local Covering for Previously Covered Inputs) is proposed, which executes Local Covering only if the input has been previously covered by the covering processes. The experimental results with three different problems show that XCS-LCPCI successfully acquires the general classifiers with any of six different parameter settings, while the performance of the conventional XCS varies depending on the parameter settings.

Model extraction (ME) attacks have been shown to cause financial losses for Machine-Learning-as-a-Service (MLaaS) providers. Attackers steal ML models on MLaaS platforms by building substitute models using queries to and responses from MLaaS platforms. The ML models targeted by attackers are called targeted models. In previous studies, researchers have assumed that attackers build substitute models that classify the same number of classes as targeted ones, which classify thousands or millions of classes to meet users' diverse expectations. We call such models general-purpose models. In fact, attackers can monetize stolen models if they accurately distinguish some classes from others. We call such models special-purpose models. For instance, a model that detects vehicles is useful for collision avoidance systems, and a model that detects wild animals is useful to drive them away from agricultural land. In this work, we investigate a threat of special-purpose ME attacks that steal special-purpose models. Our experimental results show that attackers can build an accurate special-purpose model, which achieves an 80% f-measure, with as few as 100 queries in the worst case. We discuss the difficulty in preventing the attacks with previously proposed defense methods and point out the necessity of a new defense method.

This paper proposes a novel learning classifier system (LCS) framework named ELSDeCS (Encoding, Learning, Sampling, and Decoding Classifier System) which can employ any dimensionality reduction method as pre-processing of learning and has its own components for extracting interpretable rule representations. We also propose two LCSs as examples of ELSDeCS. The first is DCAXCSR2, which is a revised version of the conventional system, and the second is VAEXCSR, which employs a deep generative model for dimensionality reduction. The experimental results on a classification task of handwritten digits show that only VAEXCSR can extract useful rule representations thanks to its robustness of decoding newly generated samples.

We present a new method of rendering aerial haptic images that uses femtosecond-laser light fields and ultrasonic acoustic fields. In conventional research, a single physical quantity has been used to render aerial haptic images. In contrast, our method combines multiple fields (light and acoustic fields) at the same time. While these fields have no direct interference, combining them provides benefits such as multi-resolution haptic images and a synergistic effect on haptic perception. We conducted user studies with laser haptics and ultrasonic haptics separately and tested their superposition. The results showed that the acoustic field affects the tactile perception of the laser haptics. We explored augmented reality/virtual reality (AR/VR) applications such as providing haptic feedback of the combination of these two methods. We believe that the results of this study contribute to the exploration of laser haptic displays and expand the expression of aerial haptic displays based on other principles.

We present a method of rendering aerial and volumetric graphics using femtosecond lasers. A high-intensity laser excites physical matter to emit light at an arbitrary three-dimensional position. Popular applications can thus be explored, especially because plasma induced by a femtosecond laser is less harmful than that generated by a nanosecond laser. There are two methods of rendering graphics with a femtosecond laser in air: producing holograms using spatial light modulation technology and scanning of a laser beam by a galvano mirror. The holograms and workspace of the system proposed here occupy a volume of up to 1 cm 3 ; however, this size is scalable depending on the optical devices and their setup. This article provides details of the principles, system setup, and experimental evaluation, and discusses the scalability, design space, and applications of this system. We tested two laser sources: an adjustable (30--100fs) laser that projects up to 1,000 pulses/s at an energy of up to 7mJ/pulse and a 269fs laser that projects up to 200,000 pulses/s at an energy of up to 50μJ/pulse. We confirmed that the spatiotemporal resolution of volumetric displays implemented using these laser sources is 4,000 and 200,000 dots/s, respectively. Although we focus on laser-induced plasma in air, the discussion presented here is also applicable to other rendering principles such as fluorescence and microbubbles in solid or liquid materials.

This paper describes a design approach and its implementation for a broadband access management system that provides element and network management functions for FTTC (fiber-to-the-curb)/FTTB (fiber-to-the-building) access networks, and service management functions for video on demand (VOD) services. The proposed architecture is a top-down approach, and consists of four levels of basic architecture, namely, service/functional architecture, system architecture, network architecture and software architecture. By the proposed architecture, TMN compliant layered management functions are realized for the broadband access system at EML (element management layer), NML (network management layer) and SML (service management layer). By considering multi-vendor environments, the proposed management systems communicate each other via OSI protocols. Three different configurations are evaluated for EML agents, and the distributed or hybrid EML agent approaches are cost effective solutions based on current technologies. Finally, this paper shows a prototype system based on the proposed architecture.

As the deployment of high-speed fiber transmission systems has been accelerated, they are widely recognized as a firm infrastructure of the information society. In SONET, the self-healing networks have been highlighted as one of the most advanced mechanisms used to realize SONET survivable networks. Several schemes have been proposed and studied due to the rapid progress in the development of highly intelligent network elements (NEs). A DCS based distributed self-healing network is discussed from the viewpoint of its control algorithms. The self-healing algorithm, called TRANS, is explained in detail. It possesses such desirable features as providing fast and flexible restoration with line and path level restoration applied to an individual STS-1 channel, the capability of handling multiple and even node failures, and so on. Both software simulation and hardware experiment verify that TRANS works properly in a real distributed environment, the result of which is shown in the paper. In addition, the combined use of TRANS and ring restoration control is proposed taking into account a practical SONET application.>

An advanced groupband data modem based on the orthogonally multiplexed QAM technique is proposed. Its performance is evaluated through experiments. In the proposed modem, data are transmitted by a number of orthogonally multiplexed parallel subchannels. Because of this orthogonally multiplexing parallelism, the proposed modem becomes more immune to Gaussian noises, impulsive noises, and channel distortions, compared to the conventional Class IV partial response VSB modem. Furthermore, the orthogonally multiplexed QAM modem retains high flexibility in multiplexing a variety of data terminals and in carrying out fallback operation to a wide range of data rates. Experimental results demonstrate the robustness of the proposed groupband data modem.