Information-Centric Networking (ICN) is a promising foundation for adaptive video streaming, enabling content-based control and in-network caching to improve Quality of Experience (QoE). Fast-Retransmission Notification Control (FNC) promotes fair bitrate adaptation by selectively notifying disadvantaged clients based on active stream bitrate levels. However, FNC applies uniform control only to low-bitrate clients, leaving high-bitrate clients unmanaged, which may cause queueing delays under congestion. This paper proposes Fast-Retransmission Notification and Explicit Cache Placement Notification (FNECPN), an FNC extension that introduces bitrate-based control by offloading high-bitrate content to router caches via Explicit Cache Placement Notification (ECPN), thereby accelerating low-bitrate delivery. Simulations on a realistic topology show that FNECPN improves fairness and responsiveness, achieving 27% higher QoE, 23% better throughput, and 40% shorter queueing delays compared to state-of-the-art ABR algorithms.

Extended Reality (XR) applications challenge network capabilities owing to their demanding requirements of high bandwidth, ultra-low latency, and reliable delivery. These applications generate heterogeneous traffic patterns, including time-sensitive control data, bandwidth-intensive video streams, and consistent audio streams, each with distinct QoS requirements. Current approaches fail to handle multiple streams with diverse QoS requirements, resulting in performance degradation under varying traffic conditions. We propose BITEX (BandWidth Trading for Extended Reality), a cross-layer enhancement mechanism for TCP BBR that intelligently manages bandwidth allocation across traffic classes using a state-based trading system. In addition, BITEXv2, which implements both data rate adjustment and TCP rate control, achieves better performance with improved fairness metrics. This cross-layer design combines transport-layer modifications with application-layer bandwidth trading through a bidirectional feedback loop, effectively addressing the fundamental limitations of single-layer solutions and enabling the reliable deployment of XR applications in dynamic network environments. Performance evaluation across AR, VR, and CG (cloud gaming) scenarios demonstrate that our proposed method outperforms standard TCP BBR in delay, packet loss, throughput stability, and QoE metrics under varying user loads.

Today, information centric networking enables adaptive video streaming clients to further improve QoE by applying flexible content-based control. However, an adaptive bitrate algorithm makes a client occupy the bottleneck link at excessively high bitrate, reducing the QoE fairness to other clients sharing the bottleneck link. Then, we propose fairAccel, a method of accelerating bitrate-based feedback control for achieving QoE fairness. fairAccel assigns more bandwidth to clients selecting the lower bitrate while suppressing content requests from clients selecting the highest bitrate on the bottleneck link. In addition, to further improve QoE fairness, fairAccel exploits the symmetric routing of ICN content request / response and applies bidirectional feedback control to the content request / response path. Thus, fairAccel accelerates feedback control by mitigating router queues under control of suppressing content requests before excessive traffic is delivered to the response path. Through simulation experiments, fairAccel improves the average bitrate and further improves QoE fairness for representative ABR algorithms.

In recent years, Internet of Things (IoT) has become an essential element of our daily lives. However, IoT devices used in IoT environments have limited available resources due to power and cost constraints, and this fact makes it difficult to implement advanced security measures on them. In fact, zero-day attacks targeting vulnerable IoT devices have occurred, and introducing an anomaly-based intrusion detection system (IDS) that can detect zero-day attacks is one of the countermeasures against the attacks. However, existing methods still suffer from limited detection ability due to a lack of training data. To solve this problem, this paper proposes an intrusion detection method that aggregates zero-day and false positive (FP) attack candidates extracted by an unsupervised anomaly detection algorithm using a one-class classification algorithm and FL. The detection performance evaluation confirms that the proposed method can share the autonomously detected zero-day attacks among IoT networks while suppressing FPs generated during the candidate extraction process.

Concurrent transmission (CT) is a promising technology for achieving efficient data transmission based on constructive interference of wireless broadcasts. However, it fails to transmit packets because of the destructive interference. Opportunistic routing (OR) avoids transmission failure by using path diversity based on flexible forwarder selection. However, it fundamentally imposes transmission delay owing to the packet-level forwarder selection. In this paper, we propose backoff-based OR with CT, which efficiently forwards data packets using CT, and then supports them based on OR. Finally, we compared the performance of the proposed method with that of traditional approaches using a simple lattice topology. The simulation results revealed that the proposed method achieved both a high packet transmission success rate and a small transmission delay while suppressing an increase in the network load.

AI processing units (APUs) may need to perform isolated execution for the confidentiality of model information and the integrity of execution tasks. Isolated execution of APUs significantly reduces the processing performance, which results in about three times the execution time overhead due to trusted execution environment (TEE) and cryptographic processing for isolated execution according to previous work. In this paper, we propose a high-performance, practical, and general-purpose isolated execution method of APUs by jointly co-designing APU data transfer and model encryption. In the evaluation results, our method can improve the execution time overhead by a factor of 2 or less compared with the previous work.

In recent years, the threat of attacks against IoT (Internet of Things) devices has become more apparent with the rapid growth of IoT technologies. Although an intrusion detection system (IDS) mitigates the risk of such attacks, the installation requires software rewriting or hardware replacement and causes cost increase of IoT devices or IoT gateways. In this paper, we propose an intrusion detection method that actively obtains feature values from IoT devices focusing on detecting an indication of intrusions with an essential cost.

MOTIVATION: Double-stranded RNAs (dsRNAs) are potent triggers of innate immune responses upon recognition by cytosolic dsRNA sensor proteins. Identification of endogenous dsRNAs helps to better understand the dsRNAome and its relevance to innate immunity related to human diseases. RESULTS: Here, we report dsRID (double-stranded RNA identifier), a machine-learning-based method to predict dsRNA regions in silico, leveraging the power of long-read RNA-sequencing (RNA-seq) and molecular traits of dsRNAs. Using models trained with PacBio long-read RNA-seq data derived from Alzheimer's disease (AD) brain, we show that our approach is highly accurate in predicting dsRNA regions in multiple datasets. Applied to an AD cohort sequenced by the ENCODE consortium, we characterize the global dsRNA profile with potentially distinct expression patterns between AD and controls. Together, we show that dsRID provides an effective approach to capture global dsRNA profiles using long-read RNA-seq data. AVAILABILITY AND IMPLEMENTATION: Software implementation of dsRID, and genomic coordinates of regions predicted by dsRID in all samples are available at the GitHub repository: https://github.com/gxiaolab/dsRID.

In recent years, network virtualization has been attracting attention since it enables diverse network construction according to user requirements, and efficiently utilizes communication resources in network infrastructures. In particular, incorporating virtual networks into a substrate network with a physical medium is an important issue in realizing the allocation of communication resources in network virtualization. This is called Virtual Network Embedding (VNE) and has been widely studied. Many algorithms have been proposed for VNE in wired networks. On the other hand, VNE for wireless networks has not been well-studied since it is difficult to abstract wireless communication resources due to problems specific to the wireless environment such as broadcastability, attenuation, and the variety of access technologies. However, with the development of multiplexing and frequency utilization technologies for wireless transmission, VNE algorithms have been studied in wireless networks in recent years. A VNE algorithm considering three QoS types was proposed in the previous study. In this study, we introduce a hybrid approach of guaranteed and best-effort forms to this VNE algorithm to improve the request acceptance rate of virtual resources and the efficiency of substrate resource usage.

Named Data Networking (NDN) is a widely adopted future Internet architecture well-suited to large scale content retrieval.The congestion control is one of the research topics actively studied, and the rate-based congestion control method is considered to be fitted to NDN.From the viewpoint of implementation, however, the rate-based method has an issue that it requires the fine-grained clock management, which is hard to implement in off-the-shelf computers.Among the rate-based congestion control methods, an approach in which intermediate nodes report a maximum rate explicitly for a flow is considered to work well.In this paper, we pick up the Multipath-aware ICN Rate-based Congestion Control as an example of explicit rate reporting scheme, and examine how coarse-grained clock gives impacts to its performance.This paper provides the performance evaluation when consumers and NDN routers use the system clock with long time interval.This paper also proposes a method for smoothening Interest sending under a coarse-grained clock and evaluates the performance of proposed method.

Floating content (FC) is a decentralized on-demand information sharing method with ad hoc communication among terminals. In FC, all contents are anchored to specific locations and shared among mobile terminals inside a predefined sharing area, which is referred to as an anchor zone (AZ), as if the contents float in the geographical area. However, contents may be vanished if there is no terminal holding the contents inside an AZ because the sharing procedure depends on ad hoc communication and infrastructure-based supports are not available. Thus, terminal density in the AZ significantly affects the content lifetime since there are less communication opportunities among terminals in sparse environments and disseminating contents are difficult. To address this issue, a method that adaptivel adjusts the effective radius of an AZ, namely the area in which information is shared with FC, has been proposed. The principle of the method is to expand the area according to the terminal density in the AZ. Although expanding the effective radius easily prolongs the lifetime of contents, unregulated expansion may lead to excessive and unnecessary dissemination. This paper proposes an effective radius adjustment and transmission timing control method to realise efficient and effective information sharing with FC. The proposed method employs terminal density and terminal encounter rate to estimate further communication opportunities. The simulation results show that the proposed method could achieve longer content lifetime with less overhead.

CCN (Content-Centric Network) has attracted attention as new network architecture suitable for content distribution. For Quality of Service (QoS) control of CCNs in previous studies, the priority has been controlled in a relative manner using caching strategies or packet scheduling in routers. However, QoS controls based on relative priority are insufficient because video content distribution requires bandwidth for the bit rate of the video content. We propose a bandwidth reservation method for the CCN architecture to prepare the required bandwidth considering the bit rate of the video content. In the proposed method, the CCN routers adaptively reserve bandwidth using the data size requested by Interest packets. The bandwidth is then prepared for users without knowing the number of users, even when they simultaneously request different parts of the same content. The proposed method is implemented in CCNx, and its effectiveness is shown experimentally.

In ad hoc networks, backoff-based OR (opportunistic routing) protocols, which autonomously selects a forwarder among the potential forwarders based on a random backoff time, have been proposed. However, each potential forwarder must wait for the backoff time to avoid packet collisions among receivers. In addition, the terminal density strongly affects the performance since the backoff-based OR protocols improve the performance by using multiple terminals. In this paper, we propose a novel OR protocol called PRIOR (prioritized forwarding for opportunistic routing). In PRIOR, a terminal called prioritized forwarder, which forwards packets without using a backoff time, is selected among neighbours. In addition, we propose a hop-by-hop retransmission control that performs the retransmission terminal selection on the basis of a neighbour relation with the PFs. Moreover, we introduce an explicit acknowledgement mechanism on the basis of the difference of hop counts to alleviate the bad effect of the retransmission control in dense environments. Finally, we evaluate PRIOR in comparison with conventional backoff-based OR protocols in computer simulation.

In ad hoc networks, radio interference and terminal' mobility can cause frame losses. When the frame losses occur continuously, general datalink layer protocols regard it as a link failure and then notify it to network layer protocols. In conventional routing protocols, a source terminal initiates route discovery process to recover from the link failures on receiving the notifications even if the source terminal still has a valid route to a destination terminal. However, the route discovery requires a route request flooding and the flooding may consume limited network resources of ad hoc networks and that also degrades transmission efficiency. For solving the problems, several retransmission and relay control methods have been proposed. In these methods, neighbour terminals of a transmission route overhear transmitted frames and retransmit lost frames autonomously by using the overheard frames. However, the conventional retransmission methods have a disadvantage that they degrade the transmission efficiency under bursty frame loss interval. Moreover, conventional relay control methods suffer from a processing overhead and a slow adaptation problem under a sparse frame loss. In this paper, we propose an efficient and lightweight method that neighbour terminals adaptively retransmit lost frames considering a frequency of the losses. Computer simulations clarify the performances and the effectiveness of the proposed method.

Recent simulation-based networking research prefers the use of real packet traces to traditional model-based traffic generation. Part of this demand comes from the fact that synthetic traffic is only remotely similar to that found in real life. Now, the biggest problem with replaying real packet traces in simulation is that the majority of publicly available traces are bilateral. Splitting traces into two subtraces for each direction of the link is a challenging task and has already been tackled by several research works and software tools with imperfect outcome. This research proposes a new method for two-side splitting of packet traces.

In this paper, we evaluate a method for on-line handwritten Kanji character recognition by describing the structure of Kanji using stochastic context-free grammar (SCFG), and extend it in order to recognize Kanji strings. In this method, we turn attention to the hierarchical structure of Kanji characters which consist of character-parts and strokes, and consider all character patterns or strings to be generated from SCFG with stochastic stroke shape and position relationship between character-parts. Describing Kanji with a few stroke shape and relative position labels, the method enables efficient training and thus robust recognition. We evaluated the recognition performance on several domains of Kanji, and on Kanji strings consist of 2 or 3 characters and gained the recognition rate of 99.29 - 97.40% for characters and 90.80% for strings.

This paper presents a new approach to online recognition of handwritten Kanji characters focusing on their hierarchical structure. Stochastic context-free grammar (SCFG) is introduced to represent the Kanji character generating process in combination with Hidden Markov Models (HMM) representing Kanji substrokes and to improve the recognition accuracy of important and frequently used Kanji characters in which inter-stroke relative positions play important roles. Combining the stroke likelihood and the relative-position likelihood between character-parts in the parsing process is expected to compensate their ambiguities. By modeling relative positions and share the models across distinct Kanji categories, a small training data can yield effective results and enables us to recognize Kanji simply by defining the SCFG rules to represent their structures without training data. Experimental results on an online handwritten Kanji database from JAIST (Japan Advanced Institute of Science and Technology) showed significant improvements in the recognition rates of some important Kanji with relatively fewer strokes and also showed little difference between the trained- and the non-trained Kanji in recognition rates.