A5071595184- Topic Modeling
- Advanced Text Analysis Techniques
- Recommender Systems and Techniques
- Natural Language Processing Techniques
- Web Data Mining and Analysis
- Anomaly Detection Techniques and Applications
- Text and Document Classification Technologies
- Advanced Graph Neural Networks
- Sentiment Analysis and Opinion Mining
- Data Stream Mining Techniques
- Semiconductor materials and devices
- Manufacturing Process and Optimization
- Time Series Analysis and Forecasting
- Business Process Modeling and Analysis
- Domain Adaptation and Few-Shot Learning
- Semantic Web and Ontologies
- Bayesian Modeling and Causal Inference
- Robotics and Automated Systems
- Speech and dialogue systems
- Caching and Content Delivery
- Music and Audio Processing
- Image Processing Techniques and Applications
- Multimodal Machine Learning Applications
- Advancements in Semiconductor Devices and Circuit Design
- Graph Theory and Algorithms
Yonsei University
2023-2025
Samsung (South Korea)
2022-2025
University of Illinois Urbana-Champaign
2000-2023
University of Minnesota System
2023
Pohang University of Science and Technology
2016-2022
Korea Post
2020-2021
University of Duisburg-Essen
2021
Upper Iowa University
2021
SK Group (South Korea)
2018
Sungkyunkwan University
2012-2018
Providing accurate recommendations to newly joined users (or potential users, so-called cold-start users) has remained a challenging yet important problem in recommender systems. To infer the preferences of such based on their observed other domains, several cross-domain recommendation (CDR) methods have been studied. The state-of-the-art Embedding and Mapping approach for CDR (EMCDR) aims latent vectors by supervised mapping from space another domain. In this paper, we propose novel...
The goal of one-class collaborative filtering (OCCF) is to identify the user-item pairs that are positively-related but have not been interacted yet, where only a small portion positive interactions (e.g., users' implicit feedback) observed. For discriminative modeling between and negative interactions, most previous work relied on sampling some extent, which refers considering unobserved as negative, actual ones unknown. However, scheme has critical limitations because it may choose...
Recent studies on electronic health records (EHRs) started to learn deep generative models and synthesize a huge amount of realistic records, in order address significant privacy issues surrounding the EHR. However, most them only focus structured about patients' independent visits, rather than chronological clinical records. In this article, we aim sequences EHRs based autoencoder.
In specialized fields like the scientific domain, constructing large-scale human-annotated datasets poses a significant challenge due to need for domain expertise. Recent methods have employed large language models generate synthetic queries, which serve as proxies actual user queries. However, they lack control over content generated, often resulting in incomplete coverage of academic concepts documents. We introduce Concept Coverage-based Query set Generation (CCQGen) framework, designed...
Recent advancements in table-based reasoning have expanded beyond factoid-level QA to address insight-level tasks, where systems should synthesize implicit knowledge the table provide explainable analyses. Although effective, existing studies remain confined scenarios a single gold is given alongside user query, failing cases users seek comprehensive insights from multiple unknown tables. To bridge these gaps, we propose MT-RAIG Bench, design evaluate on Retrieval-Augmented Insight...
ABSTRACT The effects of ozone at 0.25, 0.40, and 1.00 ppm on Listeria monocytogenes were evaluated in distilled water phosphate-buffered saline. Differences sensitivity to found exist among the six strains examined. Greater cell death was following exposure lower temperatures. Early stationary-phase cells less sensitive than mid-exponential- late cells. Ozonation cabbage inoculated with L. effectively inactivated all after 5 min. abilities vivo catalase superoxide dismutase protect from also...
With the increase of available time series data, predicting their class labels has been one most important challenges in a wide range disciplines. Recent studies on classification show that convolutional neural networks (CNN) achieved state-of-the-art performance as single classifier. In this work, pointing out global pooling layer is usually adopted by existing CNN classifiers discards temporal information high-level features, we present dynamic (DTP) technique reduces size hidden...
Recent recommender systems have shown remarkable performance by using an ensemble of heterogeneous models. However, it is exceedingly costly because requires resources and inference latency proportional to the number models, which remains bottleneck for production. Our work aims transfer knowledge teachers a lightweight student model distillation (KD), reduce huge costs while retaining high accuracy. Through empirical study, we find that efficacy severely drops when transferring from...
This study developed a smartphone application that provides wireless communication, NRTIP client, and RTK processing features, which can simplify the Network RTK-GPS system while reducing required cost. A determination method for an error model in measurements was proposed, considering both random autocorrelation errors, to accurately calculate coordinates measured by using state estimation filters. The performance evaluation of showed it could perform high-precision real-time positioning,...
Recently, web platforms are operating various service domains simultaneously. Targeting a platform that operates multiple domains, we introduce new task, Multi-Domain Recommendation to Attract Users (MDRAU), which recommends items from ``unseen'' with each user has not interacted yet, by using knowledge the user's ``seen'' domains. In this paper, point out two challenges of MDRAU task. First, there numerous possible combinations mappings seen unseen because users have usually different...
Topic taxonomies, which represent the latent topic (or category) structure of document collections, provide valuable knowledge contents in many applications such as web search and information filtering. Recently, several unsupervised methods have been developed to automatically construct taxonomy from a text corpus, but it is challenging generate desired without any prior knowledge. In this paper, we study how leverage partial incomplete) about guidance find out complete taxonomy. We propose...
Recently, finetuning a pretrained language model to capture the similarity between sentence embeddings has shown state-of-the-art performance on semantic textual (STS) task. However, absence of an interpretation method for makes it difficult explain output. In this work, we explicitly describe distance as weighted sum contextualized token distances basis transportation problem, and then present optimal transport-based measure, named RCMD; identifies leverages semantically-aligned pairs. end,...