In the context of global warming and fossil fuel depletion, electric vehicles (EVs) have become increasingly popular for reducing both carbon emissions and fossil fuel consumption. However, as the demand for EV charging power rises along with the expansion of EVs, conventional power plants require more fuel, and carbon emissions increase. This suggests that the goal of promoting EV adoption to mitigate climate change and reduce reliance on fossil fuels may face significant challenges. Therefore, there is a need to adopt renewable energy generation for EV charging stations to maximize the effectiveness of EV distribution in an eco-friendly way. This paper aims to propose an optimal renewable energy generation system for an EV charging station, with a specific focus on the use of an actual load profile for the station, the consideration of carbon emissions and economic evaluation, and the study of a specific case location in Korea. As a case study, an EV charging station in Korea was selected, and its renewable energy fractions (REF) of 0%, 25%, 50%, 75%, and 100% were considered for comparison of carbon emissions and economic evaluation with the help of HOMER software. In addition, the system with 25% REF was analyzed to find the best operating strategy considering the climate characteristics of the case site. The results show that the system configuration of PV/ESS is the most economical among all the REF cases, including PV, WT, and ESS, due to the meteorological characteristics of the site, and that the system with REF below 25% is the most optimal in economic terms and carbon emissions.

Stream processing requires integrating with background knowledge in order to become rich knowledge. As a promising approach, it is getting important to combine streaming data with linked open data. However, since linked data change dynamically, it is impossible to synchronize their distributed data sources perfectly and seamlessly. To reduce the high cost of the synchronization, the materialized views (or views) that store local copies of remote sources are used but may degrade the accuracy of stream processing. To balance response time against accuracy, recent works manage a refresh budget—that is, the limited cost allocated for updating views over remote sources. However, they fail to allocate a refresh budget and produce a low accuracy when a tight deadline is given. To solve the problem, we propose an efficient method of allocating a refresh budget to view updates. The proposed method updates views both in the background and on demand. Experimental results with real and synthetic data sets show that the proposed method achieves superiority in terms of answer staleness, resource utilization, and refresh budget usage.

Cold items are a critical problem in the recommendation domain because newly introduced items lack user–item interactions to train accurate collaborative filters (CFs). Recent studies have adopted neural networks such as MLPs and autoencoders to predict collaborative embeddings learned by CFs, using items’ side information available at the time of registration. As a generative model, diffusion models have achieved success in various fields, such as image generation and natural language processing, through their superior generative capability. This paper proposes a diffusion model as a base for cold item recommendation by generating collaborative embeddings for cold items. First, using a diffusion model with our customized predictor, we directly generate items’ collaborative embeddings conditioned on their side information. Then, a second-stage refiner, which adopts simple MLPs and dropout, is trained to calculate the final recommendation scores. The proposed method requires only the user embeddings from the existing model and the side information of cold items, making it easy to integrate into existing recommender systems. Extensive experiments on real-world datasets show that the proposed method outperforms state-of-the-art baseline methods and indicates the potential of diffusion models in cold item recommendation.

Identifying key factors and analyzing their causal relationships significantly enhance decision-making effectiveness in site selection. Although numerous studies have applied Multi-Criteria Decision-Making (MCDM) methods to site selection, these traditional approaches often overlook or inadequately represent causal interdependencies among factors. This study addresses these limitations by utilizing open data for transparency and employing Bayesian Networks (BN) as a robust probabilistic modeling alternative. BNs effectively represent complex factor interactions, capturing both causal relationships and uncertainties. Experimental evaluations demonstrate that the proposed framework effectively calculates final site suitability probabilities by explicitly considering hierarchical dependencies, offering enhanced decision-making insights.

Product attributes on e-commerce pages are often distributed across two modalities, textual content and embedded images. Such cross-modal presentation can introduce inconsistencies in wording, units, values, and attribute presence. Existing approaches match representations across modalities, but such alignment only indicates correspondence and cannot determine the relation type between paired attributes, such as equivalence and value conflict. We propose a framework that verifies attribute consistency at the attribute-pair level. First, it aligns cross-modal attributes by combining representation normalization with semantic embedding matching. Second, it classifies each aligned pair into four relation types. Third, it represents the inferred relations as typed triples in a relational graph for relational embedding learning. Finally, we evaluate the proposed framework on a real-world dataset in terms of relation classification, link prediction, and structural triple injection.

Distributed renewable energy resources (DERs) like solar and wind have become key sustainable energy sources. However, their small scale and variability challenge stable energy supply, and thus DERs should be managed collectively as virtual energy plants (VPPs). Existing energy prediction uncertainties reduce trust in VPPs, limiting market growth. To address this, we propose a trust-aware prediction adjustment method. By defining DER reliability as a quantitative metric, our approach improves prediction accuracy. Experiments on real-world data show the method outperforms existing techniques in reducing non-trust hit rates.