Given an edge-incomplete graph, how can we accurately find its missing links? The problem aims to discover the missing relations between entities when their relationships are represented as a graph. Edge-incomplete graphs are prevalent in real-world due to practical limitations, such as not checking all users when adding friends in a social network. Addressing the problem is crucial for various tasks, including recommending friends in social networks and finding references in citation networks. However, previous approaches rely heavily on the given edge-incomplete (observed) graph, making it challenging to consider the missing (unobserved) links. In this paper, we propose PULL, an accurate link prediction method based on the positive-unlabeled (PU) learning. PULL treats the observed edges in the training graph as positive examples, and the unconnected node pairs as unlabeled ones. PULL effectively prevents the link predictor from blindly trusting the observed graph by proposing latent variables for every edge, and leveraging the expected graph structure with respect to these variables. Extensive experiments on real- world datasets show that PULL consistently outperforms the baselines for predicting links in edge-incomplete graphs.

The L1/L2-triggered mobility (LTM) technique, which supports beam and cell-level mobility and can significantly reduce handover (HO) interruption time, is expected to be defined as a baseline mobility technique for 5G-Advanced and 6G. However, as the HO decision criteria in LTM changes to an instantaneous value-based L1 measurement, it is expected that the problem of HO interruption due to frequent occurrence of HO will increase. Even if LTM is introduced, there is still a weakness in commercial systems where the data rate temporarily decreases due to conservative link adaptation behavior caused by the absence of initial channel state information (CSI) immediately after HO. In this paper, we propose an improved mobility management method that can respond to various signal degradation situations by simultaneously utilizing L1 measurements and L2 filtered measurements. Additionally, we propose an early CSI acquisition scheme that can acquire the CSI of the target cell before HO execution and reports the early acquired CSI to the target cell without any delay immediately after the HO. The result of the system-level simulation shows that proposed scheme reduces HO failure ratio by about 31%, and prevents a temporary low data rate due to the absence of initial CSI.

How can we accurately recommend actions for users to control their devices at home? Action recommendation for smart home has attracted increasing attention due to its potential impact on the markets of Internet of Things (IoT). However, designing an effective action recommender system is challenging because it requires handling context correlations, considering both queried contexts and previous histories of users, and dealing with capricious intentions in history. In this work, we propose SmartSense, an accurate action recommendation method for smart home. For individual action, SmartSense summarizes its device control and temporal contexts in a self-attentive manner, to reflect the importance of the correlation between them. SmartSense then summarizes sequences considering queried contexts in a query-attentive manner to extract the query-related patterns from the sequential actions. SmartSense also transfers the commonsense knowledge from routine data to better handle intentions in action sequences. As a result, SmartSense addresses all three main challenges of action recommendation for smart home, and achieves the state-of-the-art performance giving up to 9.8% higher [email protected] than the best competitor.

How can we classify graph-structured data only with positive labels? Graph-based positive-unlabeled (PU) learning is to train a binary classifier given only the positive labels when the relationship between examples is given as a graph. The problem is of great importance for various tasks such as detecting malicious accounts in a social network, which are difficult to be modeled by supervised learning when the true negative labels are absent. Previous works for graph-based PU learning assume that the prior distribution of positive nodes is known in advance, which is not true in many real-world cases. In this work, we propose GRAB (Graph-based Risk minimization with iterAtive Belief propagation), a novel end-to-end approach for graph-based PU learning that requires no class prior. GRAB models a given graph as a Markov network and runs the marginalization and update steps iteratively. The marginalization step estimates the marginals of latent variables, while the update step trains a classifier network utilizing the computed priors in the objective function. Extensive experiments on five datasets show that GRAB achieves state-of-the-art accuracy, even compared with previous methods that are given the true prior.

Vehicle-to-Everything (V2X) communication is a key enabling factor for fully autonomous driving vehicles. To this end, the 3GPP has introduced Cellular V2X (C-V2X) standards in Release 14. For Vehicle-to-Vehicle (V2V) communication, C-V2X provides the distributed resource allocation mode, termed Mode 4, which works for sensing-based semi-persistent scheduling. However, because of the sensing-based and distributed nature, Mode 4 suffers resource collision due to congestion, channel performance degradation due to blockage, etc. Thus, making an accurate assessment of resource use in Mode 4 becomes an important issue. To address this issue, we propose a scheme for PREdictive assessment of resource usage in C-V2V Mode 4, named PRESS. In PRESS, each vehicle leverages aggregate reselection counter information to predict the channel usage status for future resource use. With the assessment of resource usage related to the transmission time, a VUE can increase the possibility of choosing the least used resources. Through simulation, we confirm that PRESS outperforms the legacy scheme in terms of packet reception ratio.

Along with the surge of data traffic amount, Long Term Evolution (LTE)-based Device-to-Device (D2D) communication is emerging as a key data traffic offloading technology. However, current LTE-based D2D communication has limitations such as the lack of feedback mechanism, causing difficulty for efficient radio resource use. In this paper, we propose a feedback mechanism as well as a feedback-based rate adaptation scheme to increase the spectral efficiency of LTE-based D2D communication. In particular, the proposed feedback mechanism between Transmitter (Tx) and Receivers (Rxs) is designed to minimize signaling overhead. Thanks to the feedback mechanism, the proposed rate adaptation scheme makes D2D Tx use the highest Modulation and Coding Scheme (MCS) level while guaranteeing reliable transmission to all Rxs. Through extensive simulations, we verify that the proposed solution achieves solid goodput performance under various channel environments.