arXiv subjects

Hee-Dong Kim

Publications and source records attributed to Hee-Dong Kim.

5 recordsLinked to original sources

arXiv2025

Local Representative Token Guided Merging for Text-to-Image Generation

Stable diffusion is an outstanding image generation model for text-to-image, but its time-consuming generation process remains a challenge due to the quadratic complexity of attention operations. Recent token merging methods improve efficiency by reducing the number of tokens during attention operations, but often overlook the characteristics of attention-based image generation models, limiting their effectiveness. In this paper, we propose local representative token guided merging (ReToM), a novel token merging strategy applicable to any attention mechanism in image generation. To merge tokens based on various contextual information, ReToM defines local boundaries as windows within attention inputs and adjusts window sizes. Furthermore, we introduce a representative token, which represents the most representative token per window by computing similarity at a specific timestep and selecting the token with the highest average similarity. This approach preserves the most salient local features while minimizing computational overhead. Experimental results show that ReToM achieves a 6.2% improvement in FID and higher CLIP scores compared to the baseline, while maintaining comparable inference time. We empirically demonstrate that ReToM is effective in balancing visual quality and computational efficiency.

cs.CV↗

arXiv2023

Pioneering EEG Motor Imagery Classification Through Counterfactual Analysis

The application of counterfactual explanation (CE) techniques in the realm of electroencephalography (EEG) classification has been relatively infrequent in contemporary research. In this study, we attempt to introduce and explore a novel non-generative approach to CE, specifically tailored for the analysis of EEG signals. This innovative approach assesses the model's decision-making process by strategically swapping patches derived from time-frequency analyses. By meticulously examining the variations and nuances introduced in the classification outcomes through this method, we aim to derive insights that can enhance interpretability. The empirical results obtained from our experimental investigations serve not only to validate the efficacy of our proposed approach but also to reinforce human confidence in the model's predictive capabilities. Consequently, these findings underscore the significance and potential value of conducting further, more extensive research in this promising direction.

eess.SP↗

arXiv2021

Counterfactual Explanation Based on Gradual Construction for Deep Networks

To understand the black-box characteristics of deep networks, counterfactual explanation that deduces not only the important features of an input space but also how those features should be modified to classify input as a target class has gained an increasing interest. The patterns that deep networks have learned from a training dataset can be grasped by observing the feature variation among various classes. However, current approaches perform the feature modification to increase the classification probability for the target class irrespective of the internal characteristics of deep networks. This often leads to unclear explanations that deviate from real-world data distributions. To address this problem, we propose a counterfactual explanation method that exploits the statistics learned from a training dataset. Especially, we gradually construct an explanation by iterating over masking and composition steps. The masking step aims to select an important feature from the input data to be classified as a target class. Meanwhile, the composition step aims to optimize the previously selected feature by ensuring that its output score is close to the logit space of the training data that are classified as the target class. Experimental results show that our method produces human-friendly interpretations on various classification datasets and verify that such interpretations can be achieved with fewer feature modification.

cs.LG↗

arXiv2016

Resistive Switching Characteristics of Al/Si3N4/p-Si MIS-Based Resistive Switching Memory Devices

In this study, we proposed and demonstrated a self-rectifying property of silicon nitride (Si3N4)-based resistive random access memory device by employing p-type silicon (p-Si) as bottom electrode. The RRAM devices consisted of Al/Si3N4/p-Si are fabricated by a low presure chemical vapor deposition and exhibited an intrinsic diode property with non-linear current-voltage (I-V) behavior. In addition, compared to conventional metal/insulator/metal (MIM) structure of Al/Si3N4/Ti RRAM cells, operating current in whole bias regions for proposed metal/insulator/semiconductor (MIS) cells has been dramatically lowered because introduced p-Si bottom electrode efficiently suppresses the current in both low and high resistive states. As a result, the results mean that by employing p-Si as bottom electrode the Si3N4-based RRAM cells can be applied to selector-free RRAM cells.

cond-mat.mtrl-sci↗

arXiv2016

Resistive Switching Phenomena of HfO2 Films Grown by MOCVD for Resistive Switching Memory Devices

The resistive switching phenomena of HfO2 films grown by metalorganic chemical vapor deposition was studied for the application of ReRAM devices. In the fabricated Pt/HfO2/TiN memory cells, the bipolar resistive switching characteristics were observed, and the set and reset states were measured to be as low as 7 uA and 4 uA, respectively, at VREAD = 1 V. Regarding the resistive switching performance, the stable RS performance was observed under 40 repetitive dc cycling test with the small variations of set/reset voltages and currents, and good retention characteristics over 105 s in both LRS and HRS. These results show the possibility of MOCVD grown HfO2 films as a promising resistive switching materials for ReRAM applications.

cond-mat.mtrl-sci↗