arXiv subjects

Sunghyun Sim

Publications and source records attributed to Sunghyun Sim.

8 recordsLinked to original sources

arXiv2026

CF-JEPA: Mask-free forward prediction with asymmetric encoder utilization for time-series representation learning

Self-supervised learning (SSL) for time-series representation learning is dominated by two paradigms: contrastive methods, which face challenges in constructing positive or negative pairs, and masking-based methods, which disrupt the temporal continuity of time-series signals. Joint-Embedding Predictive Architecture (JEPA) offers a promising alternative by predicting in representation space rather than reconstructing raw inputs. However, existing time-series JEPA variants still rely on masking and therefore inherit its continuity problem. Crop-based Forward JEPA (CF-JEPA) is proposed as an innovative mask-free framework that replaces masking with multi-horizon forward prediction: random crops serve as context views, and short-, mid-, and long-horizon future representations are predicted in the forward temporal direction, directly leveraging the inherent temporal ordering of time-series data as a learning signal. A strong asymmetry is also identified between the online encoder and the exponential moving average (EMA) target encoder, both produced from a single training run: the online encoder develops higher-rank discriminative features, while the EMA target encoder develops smoother, lower-rank temporal features. Exploiting this asymmetry, classification is routed to the online encoder and forecasting or anomaly detection to the EMA target encoder, achieving a 27% reduction in multivariate forecasting mean squared error (MSE) at no additional training cost. Across 126 University of California, Riverside (UCR) and 26 University of East Anglia (UEA) classification datasets, eight electricity transformer temperature forecasting benchmarks, and Key Performance Indicator /Yahoo anomaly detection, CF-JEPA achieves the highest average accuracy and rank on UCR and UEA among self-supervised baselines and ranks second on univariate forecasting and k-nearest neighbors-scored anomaly detection.

cs.LG↗

arXiv2026

FEATHer: Fourier-Efficient Adaptive Temporal Hierarchy Forecaster for Time-Series Forecasting

Time-series forecasting is fundamental in industrial domains like manufacturing and smart factories. As systems evolve toward automation, models must operate on edge devices (e.g., PLCs, microcontrollers) with strict constraints on latency and memory, limiting parameters to a few thousand. Conventional deep architectures are often impractical here. We propose the Fourier-Efficient Adaptive Temporal Hierarchy Forecaster (FEATHer) for accurate long-term forecasting under severe limits. FEATHer introduces: (i) ultra-lightweight multiscale decomposition into frequency pathways; (ii) a shared Dense Temporal Kernel using projection-depthwise convolution-projection without recurrence or attention; (iii) frequency-aware branch gating that adaptively fuses representations based on spectral characteristics; and (iv) a Sparse Period Kernel reconstructing outputs via period-wise downsampling to capture seasonality. FEATHer maintains a compact architecture (as few as 400 parameters) while outperforming baselines. Across eight benchmarks, it achieves the best ranking, recording 60 first-place results with an average rank of 2.05. These results demonstrate that reliable long-range forecasting is achievable on constrained edge hardware, offering a practical direction for industrial real-time inference.

cs.LG↗

arXiv2025

Rewarding Structural Conformance of Reasoning using Process Mining

Recent advances in sparse reward policy gradient methods have enabled effective reinforcement learning (RL)-based language model post-training. However, for reasoning tasks such as mathematical problem solving, binarized outcome rewards provide limited feedback on intermediate reasoning steps. While some studies have attempted to address this issue by estimating overall reasoning quality, it remains unclear whether these rewards are reliable proxies for the quality of stepwise reasoning. In this study, we consider reasoning as a structured process and propose TACReward, the reward model that can be seamlessly integrated into sparse reward policy gradient methods without additional human annotation costs or architectural modifications. TACReward aggregates stepwise structural deviations between teacher and policy reasoning using process mining techniques, producing a scalar output reward range of [0, 1] to indicate reasoning quality. Experiments on multiple mathematical reasoning benchmarks demonstrate that integrating the TACReward into sparse reward frameworks encourages the policy model to improve the structural quality of reasoning. Consequently, this leads to consistent performance improvements over existing sparse reward frameworks. Our code and checkpoints are publicly available at https://github.com/Thrillcrazyer/TACReward and https://huggingface.co/Thrillcrazyer/TACReward7B.

cs.AI↗

arXiv2024

Distributed Lag Transformer based on Time-Variable-Aware Learning for Explainable Multivariate Time Series Forecasting

Time series data is a key element of big data analytics, commonly found in domains such as finance, healthcare, climate forecasting, and transportation. In large scale real world settings, such data is often high dimensional and multivariate, requiring advanced forecasting methods that are both accurate and interpretable. Although Transformer based models perform well in multivariate time series forecasting (MTSF), their lack of explainability limits their use in critical applications. To overcome this, we propose Distributed Lag Transformer (DLFormer), a novel Transformer architecture for explainable and scalable MTSF. DLFormer integrates a distributed lag embedding and a time variable aware learning (TVAL) mechanism to structurally model both local and global temporal dependencies and explicitly capture the influence of past variables on future outcomes. Experiments on ten benchmark and real world datasets show that DLFormer achieves state of the art predictive accuracy while offering robust, interpretable insights into variable wise and temporal dynamics. These results highlight ability of DLFormer to bridge the gap between performance and explainability, making it highly suitable for practical big data forecasting tasks.

cs.LG↗

arXiv2024

Legacy Learning Strategy Based on Few-Shot Font Generation Models for Automatic Text Design in Metaverse Content

The metaverse consists of hardware, software, and content, among which text design plays a critical role in enhancing user immersion and usability as a content element. However, in languages such as Korean and Chinese that require thousands of unique glyphs, creating new text designs involves high costs and complexity. To address this, this study proposes a training strategy called Legacy Learning, which recombines and transforms structures based on existing text design models. This approach enables the generation of new text designs and improves quality without manual design processes. To evaluate Legacy Learning, it was applied to Korean and Chinese text designs. Additionally, we compared results before and after on seven state of the art text generation models. As a result, text designs generated using Legacy Learning showed over a 30% difference in Frechet Inception Distance (FID) and Learned Perceptual Image Patch Similarity (LPIPS) metrics compared to the originals, and also exhibited meaningful style variations in visual comparisons. Furthermore, the repeated learning process improved the structural consistency of the generated characters, and an OCR based evaluation showed increasing recognition accuracy across iterations, indicating improved legibility of the generated glyphs. In addition, a System Usability Scale (SUS) survey was conducted to evaluate usability among metaverse content designers and general users. The expert group recorded a score of 95.78 ("Best Imaginable"), while the non expert group scored 76.42 ("Excellent"), indicating an overall high level of usability. These results suggest that Legacy Learning can significantly improve both the production efficiency and quality of text design in the metaverse environment.

cs.HC↗

arXiv2024

Artificial Intelligence-based Smart Port Logistics Metaverse for Enhancing Productivity, Environment, and Safety in Port Logistics: A Case Study of Busan Port

The increase in global trade, the impact of COVID-19, and the tightening of environmental and safety regulations have brought significant changes to the maritime transportation market. To address these challenges, the port logistics sector is rapidly adopting advanced technologies such as big data, Internet of Things, and AI. However, despite these efforts, solving several issues related to productivity, environment, and safety in the port logistics sector requires collaboration among various stakeholders. In this study, we introduce an AI-based port logistics metaverse framework (PLMF) that facilitates communication, data sharing, and decision-making among diverse stakeholders in port logistics. The developed PLMF includes 11 AI-based metaverse content modules related to productivity, environment, and safety, enabling the monitoring, simulation, and decision making of real port logistics processes. Examples of these modules include the prediction of expected time of arrival, dynamic port operation planning, monitoring and prediction of ship fuel consumption and port equipment emissions, and detection and monitoring of hazardous ship routes and accidents between workers and port equipment. We conducted a case study using historical data from Busan Port to analyze the effectiveness of the PLMF. By predicting the expected arrival time of ships within the PLMF and optimizing port operations accordingly, we observed that the framework could generate additional direct revenue of approximately 7.3 million dollars annually, along with a 79% improvement in ship punctuality, resulting in certain environmental benefits for the port. These findings indicate that PLMF not only provides a platform for various stakeholders in port logistics to participate and collaborate but also significantly enhances the accuracy and sustainability of decision-making in port logistics through AI-based simulations.

cs.OH↗

arXiv2022

Correlation recurrent units: A novel neural architecture for improving the predictive performance of time-series data

The time-series forecasting (TSF) problem is a traditional problem in the field of artificial intelligence. Models such as Recurrent Neural Network (RNN), Long Short Term Memory (LSTM), and GRU (Gate Recurrent Units) have contributed to improving the predictive accuracy of TSF. Furthermore, model structures have been proposed to combine time-series decomposition methods, such as seasonal-trend decomposition using Loess (STL) to ensure improved predictive accuracy. However, because this approach is learned in an independent model for each component, it cannot learn the relationships between time-series components. In this study, we propose a new neural architecture called a correlation recurrent unit (CRU) that can perform time series decomposition within a neural cell and learn correlations (autocorrelation and correlation) between each decomposition component. The proposed neural architecture was evaluated through comparative experiments with previous studies using five univariate time-series datasets and four multivariate time-series data. The results showed that long- and short-term predictive performance was improved by more than 10%. The experimental results show that the proposed CRU is an excellent method for TSF problems compared to other neural architectures.

cs.LG↗

arXiv2022

Automatic Discovery of Multi-perspective Process Model using Reinforcement Learning

Process mining is a methodology for the derivation and analysis of process models based on the event log. When process mining is employed to analyze business processes, the process discovery step, the conformance checking step, and the enhancements step are repeated. If a user wants to analyze a process from multiple perspectives (such as activity perspectives, originator perspectives, and time perspectives), the above procedure, inconveniently, has to be repeated over and over again. Although past studies involving process mining have applied detailed stepwise methodologies, no attempt has been made to incorporate and optimize multi-perspective process mining procedures. This paper contributes to developing a solution approach to this problem. First, we propose an automatic discovery framework of a multi-perspective process model based on deep Q-Learning. Our Dual Experience Replay with Experience Distribution (DERED) approach can automatically perform process model discovery steps, conformance check steps, and enhancements steps. Second, we propose a new method that further optimizes the experience replay (ER) method, one of the key algorithms of deep Q-learning, to improve the learning performance of reinforcement learning agents. Finally, we validate our approach using six real-world event datasets collected in port logistics, steel manufacturing, finance, IT, and government administration. We show that our DERED approach can provide users with multi-perspective, high-quality process models that can be employed more conveniently for multi-perspective process mining.

cs.LG↗