Jooyoung Kim

Prof. Jooyoung Kim


Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
12 October 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
JKJooyoung Kim
Jooyoung Kim
KOKangrok Oh
Kangrok Oh
BOBeomseok Oh
Beomseok Oh
The perfume industry is a suitable candidate for applying advanced natural language processing techniques, yet most existing studies focus on developing fragrance design systems based on artificial intelligence advances. To meet the increasing demand for analyzing and exploiting descriptive sentences for the fragrance market, we investigate the relationship between descriptive sentences of perfumes and their notes in this paper. Our purpose for this investigation is to build a core idea for a perfume recommendation system of descriptive sentences. To accomplish this, we propose a system for perfume note estimation of descriptive sentences based on several sentence transformer models. In our leave-one-out cross-validation tests using our dataset containing 62 perfumes and 255 perfume notes, we achieved significant performance improvements (from a 37.1∼41.1% to 72.6∼79.0% hit rate with the top five items, and from a 22.1∼31.9% to a 57.3∼63.2% mean reciprocal rank) for perfume note estimation via our fine-tuning process. In addition, some qualitative examples, including query descriptions, estimated perfume notes, and the ground truth perfume notes, are presented. The proposed system improves the perfume note estimation performances using a fine-tuning process on a newly constructed dataset containing descriptive sentences of perfumes and their notes.

3 March 2025 in IEEE Sensors Journal
Verified
Journal Article
Closed access
KYKwan Yun
Kwan Yun
JPJungwoon Park
Jungwoon Park
JKJooyoung Kim
Jooyoung Kim
TKTaewook Kim
Taewook Kim
Interest in radar-based object recognition within Intelligent Transportation Systems (ITS) is surging, attributed to its robustness against environmental variations and inherent privacy-preserving capabilities. Among radar technologies, Frequency-Modulated Continuous-Wave (FMCW) radar is well-studied and excels at classifying moving objects through frequency-domain analysis and the Doppler effect. However, it struggles with classifying stationary objects. In contrast, Impulse Radio Ultra-Wideband (IR-UWB) radar transmits ultra-short pulses over a wide bandwidth, offering high temporal resolution for time-domain analysis to classify stationary objects. It can also perform frequency-domain analysis based on the Doppler effect using Pulse Repetition Frequency (PRF). Despite these advantages, time-domain analysis for stationary object classification remains limited due to data scarcity and complex signal processing. This study proposes SMOTE-Diffusion (Diffusion Model with Synthetic Minority Over-sampling Technique) to address these challenges. SMOTE-Diffusion generates high-quality time-domain data that reflect the characteristics of stationary objects, reducing data collection burdens while ensuring diverse and realistic datasets. The generated data were evaluated using quantitative metrics such as Structural Similarity Index Measure (SSIM), Mean Squared Error (MSE), and Mean Absolute Error (MAE), as well as qualitative methods like Principal Component Analysis (PCA). Applied to classification models, the augmented dataset significantly improved the average performance compared to the non-augmented data. These results demonstrate SMOTE-Diffusion’s ability to generate realistic radar data by capturing the distribution characteristics of real datasets. Therefore, by addressing the challenges associated with detecting stationary objects, it lays the foundation for developing integrated models that can classify both stationary and moving objects using IR-UWB radar.

26 June 2025 in Applied Sciences
Verified
Journal Article
Open access
PDF file
JKJooyoung Kim
Jooyoung Kim
WLWonkyung Lee
Wonkyung Lee
JPJungwoon Park
Jungwoon Park
Large-scale match data from esports games like League of Legends are stored in complex JSON files that often exceed the input token limitations of large language models (LLMs), restricting advanced analysis and applications such as automated commentary and strategic insight generation. This paper introduces the League of Legends Match Data Compactor (LoL-MDC), a tool designed to transform extensive match data into a concise and structured format optimized for LLM processing. By systematically summarizing structured match information—including match overviews, player and team statistics, timeline summaries, and algorithmically selected key events—the LoL-MDC significantly reduces the data size from approximately 80,000 tokens to under 2000 tokens while retaining analytical value. This method enables LLMs to generate coherent match summaries, analyze player performances, and identify key momentum shifts more effectively than processing raw JSON files. Additionally, the LoL-MDC integrates a winning probability metric to quantitatively enhance the selection of pivotal game events, ensuring relevance in esports analytics. Experimental evaluations demonstrate that the LoL-MDC improves data processing efficiency while maintaining critical insights. The proposed approach provides a structured and adaptable framework for applying LLMs to esports analytics and can be adapted to other competitive gaming environments, supporting AI-driven applications in match analysis, player performance evaluation, and strategic forecasting.

4 July 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
JKJooyoung Kim
Jooyoung Kim
Interactive and immersive entertainment increasingly pursues multisensory engagement, yet sensory rendering has advanced mainly in sight and sound.Smell remains underused even though scent cues heighten immersion, and a practical barrier is content: there is no reliable way to turn a scene description into a scent an olfactory display can render.We address this with text-to-scent, translating natural-language scene descriptions into producible fragrance recipes that can serve as olfactory assets for entertainment.The challenge is that large language models (LLMs) interpret evocative language well but cannot enforce the structural constraints that make a recipe producible: top--middle--base (T/M/B) architecture, weight ratios, and palette compliance.We propose a hybrid LLM-Rule framework: an LLM selects notes from a curated 80-note palette, and a deterministic rule engine, parameterized by expert perfumer metadata and tunable targets, assigns T/M/B positions, fills gaps, and computes manufacture-ready weights, so every recipe is structurally valid by construction.A 2*2 factorial comparison shows the LLM and rule engine each cover a gap the other leaves open, and a co-occurrence analysis against 21,092 real perfumes confirms that LLM-selected note combinations match real perfumery practice (positive PMI).Exposed as authoring controls, the framework supplies a tunable olfactory content-generation component for entertainment.

2 September 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
YLYuChul Lee
YuChul Lee
JKJooyoung Kim
Jooyoung Kim
HSHyeRyeon Shin
HyeRyeon Shin
Game rating analysis requires locating sparse evidence in long gameplay videos, but public rating decisions provide only bag-level labels. We present a multimodal multiple instance learning (MIL) framework that represents each video as a bag of 8-s visual, audio, and text instances. Using GRAC classification decisions, we constructed a dataset of 663 continuous gameplay excerpts selected to include potentially relevant evidence while retaining ordinary gameplay. Six modalities were evaluated with five-fold cross-validation using separate training, validation, and test sets. The validation-selected pipeline achieved a mean test macro-averaged area under the receiver operating characteristic curve (macro-AUC) of $0.869 \pm 0.074$. In a descriptive comparison of ensemble configurations using all six modalities, weighted soft voting attained $0.875 \pm 0.072$. Multimodal fusion improved bag-level Content Descriptor classification. Attention-based instance relevance scores support exploratory temporal analysis, but localization was not independently validated.

1 September 2026 in IEEE Sensors Journal
Verified
Journal Article
Closed access
JPJunga Park
Junga Park
JKJooyoung Kim
Jooyoung Kim
TKTaewook Kim
Taewook Kim