Since successful city branding plays a crucial role in establishing a city’s competitiveness and uniqueness, cities worldwide are actively involved in shaping their city images. City images are multifaceted and dynamic, influenced not only by geographic, symbolic, and cultural elements individually, but also by the intricate interplay among diverse components that constitute a city. In addition, the advent of mobile devices and the rise of social media have intensified the trend of creating new city images while some existing ones fade away. However, traditional survey-based analytical methods struggle to accurately capture the complexity and temporal changes in city images. Although some researchers have explored the analysis of city images using machine learning and social media data, there is still a gap in research that comprehensively captures the intricate nature of city images, encompassing diverse themes that evolve over time. This article aims to provide a deeper understanding of the multiple characteristics and their temporal changes in city images by leveraging geographic information from social media posts. Through an extensive analysis of three popular tourist destinations in South Korea—Gongju, Gyeongju, and Jeju—we uncover the interconnections, transformations, and underlying reasons behind actual city images. Specifically, our findings reveal that Gongju is evolving from a historically focused city into a multifaceted image incorporating welfare, business, and experiential activities. Gyeongju is shifting from a traditional cultural tourism destination to one increasingly centered on leisure and culinary experiences. Jeju Island, once dominated by traditional themes such as tangerine farming and women divers, is now characterized by modern attractions such as the Aewol café street, reflecting its transition into a premier leisure destination. Furthermore, our study contributes to the validation and enrichment of existing knowledge on dynamically evolving city images by highlighting the similarities and differences between our findings and those derived from conventional survey-based approaches.

Sign language translation (SLT) is a challenging task that involves translating sign language images into spoken language. For SLT models to perform this task successfully, they must bridge the modality gap and identify subtle variations in sign language components to understand their meanings accurately. To address these challenges, we propose a novel gloss-free SLT framework called Multimodal Sign Language Translation (MMSLT), which leverages the representational capabilities of off-the-shelf multimodal large language models (MLLMs). Specifically, we use MLLMs to generate detailed textual descriptions of sign language components. Then, through our proposed multimodal-language pre-training module, we integrate these description features with sign video features to align them within the spoken sentence space. Our approach achieves state-of-the-art performance on benchmark datasets PHOENIX14T and CSL-Daily, highlighting the potential of MLLMs to be utilized effectively in SLT. Code is available at https://github.com/hwjeon98/MMSLT.

This study investigates effective feedback mechanisms to maintain human engagement in interactive machine learning (IML) systems, focusing on social media platforms. We developed “Loop,” an IML system based on human-in-the-loop (HITL) principles that recommends content while encouraging users to report inaccuracies for model refinement. Loop implements three types of artificial intelligence (AI) feedback on user reports: (a) machine learning (ML)-centric, (b) personal-centric, and (c) community-centric feedback. In addition, we evaluated the relative effectiveness of these feedback types under two different task criticality scenarios: high and low. A user study with 30 participants was conducted to evaluate Loop through questionnaires and interviews. Results showed that participants preferred algorithmic improvements for personal benefit over altruistic contributions to the community, especially for low-criticality tasks. Furthermore, personal-centric feedback had a significant impact on user engagement and satisfaction. Our findings provide insights into the effectiveness of machine feedback in HITL-ML systems, contributing to the design of more engaging and effective IML interfaces. We discuss implications and strategies for encouraging proactive user engagement in HITL-ML-based systems, emphasizing the importance of tailored feedback mechanisms.

Identifying cohesive subgraphs within networks is a fundamental problem in graph theory, relevant to various domains. The traditional clique problem, which finds fully connected subgraphs, often faces limitations due to its strict connectivity requirements. This paper introduces a novel degree-based relaxation model called Flexi-clique, where the degree constraint is adjusted sub-linearly based on the subgraph size. We establish that the maximum Flexi-clique problem is NP-hard and propose an efficient and effective peeling algorithm to address it. Our extensive experimental evaluation of real-world datasets demonstrates the effectiveness and efficiency of our approach in discovering large, cohesive subgraphs in networks.

Knowledge distillation (KD) aims to transfer knowledge from larger (teacher) to smaller (student) networks. Previous studies focus on point-to-point or pairwise relationships in embedding features as knowledge and struggle to efficiently transfer relationships of complex latent spaces. To tackle this issue, we propose a novel KD method called TopKD, which considers the global topology of the latent spaces. We define global topology knowledge using the persistence diagram (PD) that captures comprehensive geometric structures such as shape of distribution, multiscale structure and connectivity, and the topology distillation loss for teaching this knowledge. To make the PD transferable within reasonable computational time, we employ approximated persistence images of PDs. Through experiments, we support the benefits of using global topology as knowledge and demonstrate the potential of TopKD. Code is available at https://github.com/jekim5418/TopKD

COVID-19, a highly contagious global epidemic, has prompted governments to actively recommend vaccination as a crucial measure to overcome its impact. However, vaccine hesitancy remains a significant challenge, stemming from concerns related to rapid vaccine development, streamlined clinical trials, misinformation, and potential side effects. To address these concerns, an in-depth understanding of COVID-19 vaccine side effects is paramount. This article aims to analyze COVID-19 vaccine side effects using machine learning applied to Twitter, a representative social media platform. Thorough experiments show that we can not only detect officially known COVID-19 vaccine side effects, such as pain and headache but also identify previously unknown COVID-19 vaccine side effects like myocarditis and thrombosis. More importantly, we show that connectivity analysis and cluster analysis can provide a more detailed understanding of vaccine side effects, including differences from conventional text-mining analysis results. This article has the potential to alleviate public anxiety by discovering and analyzing vaccine side effects through social media data analysis. In addition, the proposed method is more important because it can be applied not only to COVID-19 vaccines but also to other side effects related to other medications.

The creative web is all about combining different types of media to create a unique and engaging online experience. Multimodal data, such as text and images, is a key component in the creative web. Social media posts that incorporate both text descriptions and images offer a wealth of information and context. Text in social media posts typically relates to one topic, while images often convey information about multiple topics due to the richness of visual content. Despite this potential, many existing multimodal topic models do not take these criteria into account, resulting in poor quality topics being generated. Therefore, we proposed a Coherent Topic modeling for Multimodal Data (CTM-MM), which takes into account that text in social media posts typically relates to one topic, while images can contain information about multiple topics. Our experimental results show that CTM-MM outperforms traditional multimodal topic models in terms of classification and topic coherence.

Short texts are a common source of knowledge, and the extraction of such valuable information is beneficial for several purposes. Traditional topic models are incapable of analyzing the internal structural information of topics. They are mostly based on the co-occurrence of words at the document level and are often unable to extract semantically relevant topics from short text datasets due to their limited length. Although some traditional topic models are sensitive to word order due to the strong sparsity of data, they do not perform well on short texts. In this paper, we propose a novel word embedding-based topic model (WETM) for short text documents to discover the structural information of topics and words and eliminate the sparsity problem. Moreover, a modified collapsed Gibbs sampling algorithm is proposed to strengthen the semantic coherence of topics in short texts. WETM extracts semantically coherent topics from short texts and finds relationships between words. Extensive experimental results on two real-world datasets show that WETM achieves better topic quality, topic coherence, classification, and clustering results. WETM also requires less execution time compared to traditional topic models.

BACKGROUND: Text mining in the biomedical field has received much attention and regarded as the important research area since a lot of biomedical data is in text format. Topic modeling is one of the popular methods among text mining techniques used to discover hidden semantic structures, so called topics. However, discovering topics from biomedical data is a challenging task due to the sparsity, redundancy, and unstructured format. METHODS: In this paper, we proposed a novel multiple kernel fuzzy topic modeling (MKFTM) technique using fusion probabilistic inverse document frequency and multiple kernel fuzzy c-means clustering algorithm for biomedical text mining. In detail, the proposed fusion probabilistic inverse document frequency method is used to estimate the weights of global terms while MKFTM generates frequencies of local and global terms with bag-of-words. In addition, the principal component analysis is applied to eliminate higher-order negative effects for term weights. RESULTS: Extensive experiments are conducted on six biomedical datasets. MKFTM achieved the highest classification accuracy 99.04%, 99.62%, 99.69%, 99.61% in the Muchmore Springer dataset and 94.10%, 89.45%, 92.91%, 90.35% in the Ohsumed dataset. The CH index value of MKFTM is higher, which shows that its clustering performance is better than state-of-the-art topic models. CONCLUSION: We have confirmed from results that proposed MKFTM approach is very efficient to handles to sparsity and redundancy problem in biomedical text documents. MKFTM discovers semantically relevant topics with high accuracy for biomedical documents. Its gives better results for classification and clustering in biomedical documents. MKFTM is a new approach to topic modeling, which has the flexibility to work with a variety of clustering methods.

We present a method for learning dynamics of complex physical processes described by time-dependent nonlinear partial differential equations (PDEs). Our particular interest lies in extrapolating solutions in time beyond the range of temporal domain used in training. Our choice for a baseline method is physics-informed neural network (PINN) because the method parameterizes not only the solutions, but also the equations that describe the dynamics of physical processes. We demonstrate that PINN performs poorly on extrapolation tasks in many benchmark problems. To address this, we propose a novel method for better training PINN and demonstrate that our newly enhanced PINNs can accurately extrapolate solutions in time. Our method shows up to 72% smaller errors than state-of-the-art methods in terms of the standard L2-norm metric.

As location-based services using mobile devices have become globally popular these days, social network analysis (especially, community detection) increasingly benefits from combining social relationships with geographic preferences. In this regard, this article addresses the emerging problem of geosocial community detection. We first formalize the problem of geosocial co-clustering , which co-clusters the users in social networks and the locations they visited. Geosocial co-clustering detects higher-quality communities than existing approaches by improving the mapping clusterability , whereby users in the same community tend to visit locations in the same region. While geosocial co-clustering is soundly formalized as non-negative matrix tri-factorization , conventional matrix tri-factorization algorithms suffer from a significant computational overhead when handling large-scale datasets. Thus, we also develop an efficient framework for geosocial co-clustering, called GEOsocial COarsening and DEcomposition (GEOCODE) . To achieve efficient matrix tri-factorization, GEOCODE reduces the numbers of users and locations through coarsening and then decomposes the single whole matrix tri-factorization into a set of multiple smaller sub-matrix tri-factorizations. Thorough experiments conducted using real-world geosocial networks show that GEOCODE reduces the elapsed time by 19–69 times while achieving the accuracy of up to 94.8% compared with the state-of-the-art co-clustering algorithm. Furthermore, the benefit of the mapping clusterability is clearly demonstrated through a local expert recommendation application.

This paper proposes LinkBlackHole*, a novel algorithm for finding communities that are (i) overlapping in nodes and (ii) mixing (not separating clearly) in links. There has been a small body of work in each category, but this paper is the first one that addresses both. For this purpose, LinkBlackHole* incorporates the advantages of both the link-space transformation and the black hole transformation. Thorough experiments show superior quality of the communities detected by LinkBlackHole* to those detected by other state-of-the-art algorithms.

This paper proposes LinkBlackHole*, a novel algorithm for finding communities that are (i) overlapping in nodes and (ii) mixing (not separating clearly) in links. There has been a small body of work in each category, but this paper is the first one that addresses both. LinkBlackHole* is a merger of our earlier two algorithms, LinkSCAN* and BlackHole, inheriting their advantages in support of highly-mixed overlapping communities. The former is used to handle overlapping nodes, and the latter to handle mixing links in finding communities. Like LinkSCAN and its more efficient variant LinkSCAN*, this paper presents LinkBlackHole and its more efficient variant LinkBlackHole*, which reduces the number of links through random sampling. Thorough experiments show superior quality of the communities detected by LinkBlackHole* and LinkBlackHole to those detected by other state-of-the-art algorithms. In addition, LinkBlackHole* shows high resilience to the link sampling effect, and its running time scales up almost linearly with the number of links in a network.

Amulti-layer graphconsists of multiple layers of weighted graphs, where the multiple layers represent the different aspects of relationships. Considering multiple aspects (i.e., layers) together is essential to achieve a comprehensive and consolidated view. In this article, we propose a novel framework ofdifferential flattening, which facilitates the analysis of multi-layer graphs, and apply this framework to community detection. Differential flattening merges multiple graphs into a single graph such that the graph structure with the maximum clustering coefficient is obtained from the single graph. It has two distinct features compared with existing approaches. First, dealing with multiple layers is doneindependentlyof a specific community detection algorithm, whereas previous approaches rely on a specific algorithm. Thus, any algorithm for a single graph becomes applicable to multi-layer graphs. Second, the contribution of each layer to the single graph is determinedautomaticallyfor the maximum clustering coefficient. Since differential flattening is formulated by an optimization problem, the optimal solution is easily obtained by well-known algorithms such as interior point methods. Extensive experiments were conducted using the Lancichinetti-Fortunato-Radicchi (LFR) benchmark networks as well as the DBLP, 20 Newsgroups, and MIT Reality Mining networks. The results show that our approach of differential flattening leads to discovery of higher-quality communities than baseline approaches and the state-of-the-art algorithms.

In many embedded systems, particularly those with high data computations, the delay of memory access is one of the major bottlenecks in the system's performance. It has been known that there are high variations in memory-access delays depending on the ways of designing memory configurations and assigning arrays to memories. Furthermore, embedded-DRAM technology that provides efficient access modes is actively being developed, possibly becoming a mainstream in future embedded-system design. In that context, in this paper, the authors propose an effective solution to the problem of (embedded DRAM) memory allocation and mapping in memory-access-code generation with the objective of minimizing the total memory-access time. Specifically, the proposed approach, called memory-access-code optimization (MACCESS-opt), solves the three problems simultaneously: 1) determination of memories; 2) mapping of arrays to memories; and 3) scheduling of memory-access operations, so that the use of DRAM-access modes is maximized while satisfying the storage size constraint of embedded systems. Experimental data on a set of benchmark designs are provided to show the effectiveness of the proposed integrated approach. In short, MACCESS-opt reduces the total memory-access latency by over 18%, from which the authors found that the memory mapping and scheduling techniques in MACCESS-opt contribute about 12% and 6% reductions of the total memory-access latency, respectively

In many computer systems with large data computations, the delay of memory access is one of the major performance bottlenecks. In this paper, we propose an enhanced field remapping scheme for dynamically allocated structures in order to provide better locality than conventional field lay outs. Our proposed scheme reduces cache miss rates drastically by aggregating and grouping fields from multiple instances of the same structure, which implies the performance improvement and power reduction. Our methodology will become more important in the design space exploration, especially as the embedded systems for data oriented application become prevalent. Experimental results show that average L1 and L2 data cache misses are reduced by 23% and 17%, respectively. Due to the enhanced localities, our remapping achieves 13% faster execution time on average than original programs. It also reduces power consumption by 18% for data cache.

In many of embedded systems, particularly for those with high data computations, the delay of memory access is one of the major bottlenecks in the system's performance. It has been known that there are high variations in memory access delays depending on the ways of designing memory configurations and assigning arrays to memories. Furthermore, embedded DRAM technology that provides efficient access modes is actively developed, possibly becoming a mainstream in future embedded system design. In that context, in this paper we propose an effective solution to the problem of (embedded DRAM) memory allocation and mapping in memory access code generation with the objective of minimizing the total memory access time. Specifically, the proposed approach, called MACCESS-opt, solves the three problems simultaneously: (i) determination of memories, (ii) mapping of arrays to memories, and (iii) scheduling of memory access operations, so that the use of DRAM access modes is maximized while satisfying the storage size constraint of embedded system. Experimental data on a set of benchmark designs are provided to show the effectiveness of the proposed integrated approach. In short, MACCESS-opt reduces the total memory access latency by over 18%, from which we found that our memory mapping and scheduling techniques in MACCESS-opt contribute about 12% and 6% reductions of total memory access latency, respectively.