Young-Woong Ko

Domain Generalization (DG) seeks to transfer knowledge from multiple source domains to unseen target domains, even in the presence of domain shifts. Achieving effective generalization typically requires a large and diverse set of labeled source data to learn robust representations that can generalize to new, unseen domains. However, obtaining such high-quality labeled data is often costly and labor-intensive, limiting the practical applicability of DG. To address this, we investigate a more practical and challenging problem: semi-supervised domain generalization (SSDG) under a label-efficient paradigm. In this paper, we propose a novel method, CAT, which leverages semi-supervised learning with limited labeled data to achieve competitive generalization performance under domain shifts. Our method addresses key limitations of previous approaches, such as reliance on fixed thresholds and sensitivity to noisy pseudo-labels. CAT combines adaptive thresholding with noisy label refinement techniques, creating a straightforward yet highly effective solution for SSDG tasks. Specifically, our approach uses flexible thresholding to generate high-quality pseudo-labels with higher class diversity while refining noisy pseudo-labels to improve their reliability. Extensive experiments on multiple benchmark datasets demonstrate the superior performance of our method, with improvements of 3.45% on PACS, 9.47% on OfficeHome, and 10.90% on miniDomainNet datasets, highlighting its effectiveness in achieving robust generalization under domain shifts.

Deep neural networks (DNNs) have proven explicit contributions in making autonomous driving cars and related tasks such as semantic segmentation, motion tracking, object detection, sensor fusion, and planning. However, in challenging situations, DNNs are not generalizable because of the inherent domain shift due to the nature of training under the i.i.d. assumption. The goal of semantic segmentation is to preserve information from a given image into multiple meaningful categories for visual understanding. Particularly for semantic segmentation, pixel-wise annotation is extremely costly and not always feasible. Domain generalization for semantic segmentation aims to learn pixel-level semantic labels from multiple source domains and generalize to predict pixel-level semantic labels on multiple unseen target domains. In this survey, for the first time, we present a comprehensive review of DG for semantic segmentation. we present a comprehensive summary of recent works related to domain generalization in semantic segmentation, which establishes the importance of generalizing to new environments of segmentation models. Although domain adaptation has gained more attention in segmentation tasks than domain generalization, it is still worth unveiling new trends that are adopted from domain generalization methods in semantic segmentation. We cover most of the recent and dominant DG methods in the context of semantic segmentation and also provide some other related applications. We conclude this survey by highlighting the future directions in this area.

Electrocardiography (ECG)-based arrhythmia classification intends to have a massive role in cardiovascular disease monitoring and early diagnosis. However, ECG datasets are mostly imbalanced and have regularization to use real-time patient data due to privacy concerns. Traditional models do not generalize on unseen cases and are also unable to preserve data privacy. Which incentivizes performance degradation in existing models with privacy limitations. To tackle generalization and privacy issues together, we introduce the framework SF-ECG, a source-free domain adaptation approach for patient-specific ECG classification. This framework does not require source data during adaptation, which solves the privacy issue during adaptation. We adopt a generative model (GAN) that learns to synthesize patient-specific ECG data in data-inefficient classes to make additional source data for imbalanced classes. Then, we use the local structure clustering method to strongly align target ECG features with similar neighbors. After seizing clustered target features, we use a classifier that is trained on source data with generated source samples, which makes the model generalizable in classifying unseen data. Empirical results under different experimental conditions in various interdomain datasets prove that the proposed framework achieves 0.8% improvements in UDA settings, along with preserving privacy and generalizability.

본 논문에서는 데이터 수정을 최소화하기 위하여 가변 블록 파일 시스템을 제안한다. 가변 블록은 블록 기반의 스토리지에 블록의 크기보다 작은 용량의 데이터를 저장할 수 있도록 파일 시스템의 메타데이터를 수정하여 만든 개념이다. 본 논문에서 제안하는 시스템의 유용성을 보이기 위해서 사용자 수준의 파일 시스템을 구현하였다. 실험 결과 데이터 읽기와 쓰기에 대해서 기존 시스템에 비해서 제안하는 가변 블록 기법이 높은 성능을 보이는 것을 확인하였다. 1MB 파일을 대상으로 파일 수정에 대한 실험을 수행한 결과 50배의 성능 향상을 보였다. 또한 1MB파일에서 가변 블록을 최대한 모두 사용하여 데이터 수정을 하였을 때에 기존 파일 시스템에 비해 10배 이상 빠른 수행 속도를 보였다.

A bstract: Cardiovascular disease is now one of the leading causes of morbidity and mortality in humans. Electrocardiogram (ECG) is a reliable tool for monitoring the health of the cardiovascular system. Currently, there has been a lot of focus on accurately categorizing heartbeats. There is a high demand on automatic ECG classification systems to assist medical professionals. In this paper we proposed a new deep learning method called HeartNet for developing an automatic ECG classifier. The proposed deep learning method is compressed by multi-head attention mechanism on top of CNN model. The main challenge of insufficient data label is solved by adversarial data synthesis adopting generative adversarial network (GAN) with generating additional training samples. It drastically improves the overall performance of the proposed method by 5-10% on each insufficient data label category. We evaluated our proposed method utilizing MIT-BIH dataset. Our proposed method has shown 99.67 ± 0.11 accuracy and 89.24 ± 1.71 MCC trained with adversarial data synthesized dataset. However, we have also utilized two individual datasets such as Atrial Fibrillation Detection Database and PTB Diagnostic Database to see the performance of our proposed model on ECG classification. The effectiveness and robustness of proposed method are validated by extensive experiments, comparison and analysis. Later on, we also highlighted some limitations of this work.

Recent deep learning models succeed in achieving high accuracy and fast inference time, but they require high-performance computing resources because they have a large number of parameters. However, not all systems have high-performance hardware. Sometimes, a deep learning model needs to be run on edge devices such as IoT devices or smartphones. On edge devices, however, limited computing resources are available and the amount of computation must be reduced to launch the deep learning models. Pruning is one of the well-known approaches for deriving light-weight models by eliminating weights, channels or filters. In this work, we propose “zero-keep filter pruning” for energy-efficient deep neural networks. The proposed method maximizes the number of zero elements in filters by replacing small values with zero and pruning the filter that has the lowest number of zeros. In the conventional approach, the filters that have the highest number of zeros are generally pruned. As a result, through this zero-keep filter pruning, we can have the filters that have many zeros in a model. We compared the results of the proposed method with the random filter pruning and proved that our method shows better performance with many fewer non-zero elements with a marginal drop in accuracy. Finally, we discuss a possible multiplier architecture, zero-skip multiplier circuit, which skips the multiplications with zero to accelerate and reduce energy consumption.

Case based reasoning is a knowledge discovery technique that uses similar past problems to solve current new problems. It has been applied to many tasks, including the prediction of temporal variables as well as learning techniques such as neural networks, genetic algorithms, decision trees, etc. This paper presents a geometric criterion for selecting similar cases that serve as an exemplar for the target. The proposed technique, called geometric Case Based Reasoning, uses a shape distance method that uses the number of sign changes of features for the target case, especially when extracting nearest neighbors. Thus, this method overcomes the limitation of conventional case-based reasoning in that it uses Euclidean distance and does not consider how nearest neighbors are similar to the target case in terms of changes between previous and current features in a time series. These concepts are investigated against the backdrop of a practical application involving the prediction of a stock market index. The results show that the proposed technique is significantly better than the random walk model at p < 0.01. However, it was not significantly better than the conventional CBR model in the hit rate measure and did not surpass the conventional CBR in the mean absolute percentage error.