检索结果-内蒙古大学图书馆

Regularizing Hard Examples Improves Adversarial Robustness

JOURNAL OF MACHINE LEARNING RESEARCH 2025年 26卷

作者： Lee, Hyungyu Lee, Saehyung Bae, Ho Yoon, Sungroh Seoul Natl Univ Elect & Comp Engn Interdisciplinary Program Artificial Intelligence Seoul 08826 South Korea Ewha Womans Univ Dept Cyber Secur Seoul 03760 South Korea Seoul Natl Univ Interdisciplinary Program Artificial Intelligence Elect & Comp Engn AIISASRIINMC Seoul 08826 South Korea Seoul Natl Univ ISRC Seoul 08826 South Korea

Recent studies have validated that pruning hard-to-learn examples from training improves the generalization performance of neural networks (NNs). In this study, we investigate this intriguing phenomenon-the negative effect of hard examples on generalization-in adversarial training. Particularly, we theoretically demonstrate that the increase in the difficulty of hard examples in adversarial training is significantly greater than the increase in the difficulty of easy examples. Furthermore, we verify that hard examples are only fitted through memorization of the label in adversarial training. We conduct both theoretical and empirical analyses of this memorization phenomenon, showing that pruning hard examples in adversarial training can enhance the model's robustness. However, the challenge remains in finding the optimal threshold for removing hard examples that degrade robustness performance. Based upon these observations, we propose a new approach, difficulty proportional label smoothing (DPLS), to adaptively mitigate the negative effect of hard examples, thereby improving the adversarial robustness of NNs. Notably, our experimental result indicates that our method can successfully leverage hard examples while circumventing the negative effect.

关键词： adversarial robustness adversarial examples hard examples memorization robust overfitting

来源：评论

学校读者我要写书评

暂无评论

ReMP: Reusable Motion Prior for Multi-domain 3D Human Pose Estimation and Motion Inbetweening

ReMP: Reusable Motion Prior for Multi-domain 3D Human Pose E...

引用

2025 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2025

作者： Jang, Hojun Kim, Young Min Seoul National University Dept. of Electrical and Computer Engineering Korea Republic of Interdisciplinary Program in Artificial Intelligence and Inmc Seoul National University Korea Republic of

ISBN: (纸本)9798331510831

We present Reusable Motion prior (ReMP), an effective motion prior that can accurately track the temporal evolution of motion in various downstream tasks. Inspired by the success of foundation models, we argue that a robust spatio-temporal motion prior can encapsulate underlying 3D dynamics applicable to various sensor modalities. We learn the rich motion prior from a sequence of complete parametric models of posed human body shape. Our prior can easily estimate poses in missing frames or noisy measurements despite significant occlusion by employing a temporal attention mechanism. More interestingly, our prior can guide the system with incomplete and challenging input measurements to quickly extract critical information to estimate the sequence of poses, significantly improving the training efficiency for mesh sequence recovery. ReMP consistently outperforms the baseline method on diverse and practical 3D motion data, including depth point clouds, LiDAR scans, and IMU sensor data. Project page is available in https://***/ReMP. © 2025 IEEE.

关键词： Reusability

来源：评论

学校读者我要写书评

暂无评论

ReMP: Reusable Motion Prior for Multi-domain 3D Human Pose Estimation and Motion Inbetweening

ReMP: Reusable Motion Prior for Multi-domain 3D Human Pose E...

引用

IEEE Workshop on Applications of Computer Vision (WACV)

作者： Hojun Jang Young Min Kim Dept. of Electrical and Computer Engineering Seoul National University Interdisciplinary Program in Artificial Intelligence and INMC Seoul National University

ISBN: (数字)9798331510831

ISBN: (纸本)9798331510848

关键词： Training Point cloud compression Three-dimensional displays Accuracy Tracking Shape Pose estimation Noise Parametric statistics Noise measurement

来源：评论

学校读者我要写书评

暂无评论

A comprehensive survey of deep learning for time series forecasting: architectural diversity and open challenges

引用

artificial intelligence REVIEW 2025年第7期58卷 1-95页

作者： Kim, Jongseon Kim, Hyungjoon Kim, Hyungi Lee, Dongjun Yoon, Sungroh Seoul Natl Univ Interdisciplinary Program Artificial Intelligence Seoul South Korea Seoul Natl Univ Dept Elect & Comp Engn Seoul South Korea Seoul Natl Univ AIIS ASRI Seoul South Korea Seoul Natl Univ INMC Seoul South Korea R&D Dept LG Chem Seoul South Korea Samsung SDI R&D Dept Yongin South Korea

Time series forecasting is a critical task that provides key information for decision-making across various fields, such as economic planning, supply chain management, and medical diagnosis. After the use of traditional statistical methodologies and machine learning in the past, various fundamental deep learning architectures such as MLPs, CNNs, RNNs, and GNNs have been developed and applied to solve time series forecasting problems. However, the structural limitations caused by the inductive biases of each deep learning architecture constrained their performance. Transformer models, which excel at handling long-term dependencies, have become significant architectural components for time series forecasting. However, recent research has shown that alternatives such as simple linear layers can outperform Transformers. These findings have opened up new possibilities for using diverse architectures, ranging from fundamental deep learning models to emerging architectures and hybrid approaches. In this context of exploration into various models, the architectural modeling of time series forecasting has now entered a renaissance. This survey not only provides a historical context for time series forecasting but also offers comprehensive and timely analysis of the movement toward architectural diversification. By comparing and re-examining various deep learning models, we uncover new perspectives and present the latest trends in time series forecasting, including the emergence of hybrid models, diffusion models, Mamba models, and foundation models. By focusing on the inherent characteristics of time series data, we also address open challenges that have gained attention in time series forecasting, such as channel dependency, distribution shift, causality, and feature extraction. This survey explores vital elements that can enhance forecasting performance through diverse approaches. These contributions help lower entry barriers for newcomers by providing a systematic underst

关键词： Time series forecasting Deep learning Foundation model Distribution shift Causality

来源：评论

学校读者我要写书评

暂无评论

Learning 3D Scene Analogies with Neural Contextual Scene Maps

arXiv

引用

arXiv 2025年

作者： Kim, Junho Bae, Gwangtak Lee, Eun Sun Kim, Young Min Dept. of Electrical and Computer Engineering Seoul National University Korea Republic of Interdisciplinary Program in Artificial Intelligence and INMC Seoul National University Korea Republic of

Understanding scene contexts is crucial for machines to perform tasks and adapt prior knowledge in unseen or noisy 3D environments. As data-driven learning is intractable to comprehensively encapsulate diverse ranges of layouts and open spaces, we propose teaching machines to identify relational commonalities in 3D spaces. Instead of focusing on point-wise or object-wise representations, we introduce 3D scene analogies, which are smooth maps between 3D scene regions that align spatial relationships. Unlike well-studied single instance-level maps, these scene-level maps smoothly link large scene regions, potentially enabling unique applications in trajectory transfer in AR/VR, long demonstration transfer for imitation learning, and context-aware object rearrangement. To find 3D scene analogies, we propose neural contextual scene maps, which extract descriptor fields summarizing semantic and geometric contexts, and holistically align them in a coarse-to-fine manner for map estimation. This approach reduces reliance on individual feature points, making it robust to input noise or shape variations. Experiments demonstrate the effectiveness of our approach in identifying scene analogies and transferring trajectories or object placements in diverse indoor scenes, indicating its potential for robotics and AR/VR applications. © 2025, CC BY.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Battling the Non-stationarity in Time Series Forecasting via Test-time Adaptation

arXiv

引用

arXiv 2025年

作者： Kim, HyunGi Kim, Siwon Mok, Jisoo Yoon, Sungroh Department of Electrical and Computer Engineering Seoul National University Korea Republic of Interdisciplinary Program in Artificial Intelligence Seoul National University Korea Republic of AIIS ASRI INMC Seoul National University Korea Republic of

Deep Neural Networks have spearheaded remarkable advancements in time series forecasting (TSF), one of the major tasks in time series modeling. Nonetheless, the non-stationarity of time series undermines the reliability of pre-trained source time series forecasters in mission-critical deployment settings. In this study, we introduce a pioneering test-time adaptation framework tailored for TSF (TSF-TTA). TAFAS, the proposed approach to TSF-TTA, flexibly adapts source forecasters to continuously shifting test distributions while preserving the core semantic information learned during pre-training. The novel utilization of partially-observed ground truth and gated calibration module enables proactive, robust, and model-agnostic adaptation of source forecasters. Experiments on diverse benchmark datasets and cutting-edge architectures demonstrate the efficacy and generality of TAFAS, especially in long-term forecasting scenarios that suffer from significant distribution shifts. The code is available at https://***/kimanki/TAFAS. © 2025, CC BY-NC-SA.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Efficient Attention-Sharing Information Distillation Transformer for Lightweight Single Image Super-Resolution

arXiv

引用

arXiv 2025年

作者： Park, Karam Soh, Jae Woong Cho, Nam Ik Department of ECE INMC Seoul National University Seoul Korea Republic of Department of EECS Gwangju Institute of Science and Technology Gwangju Korea Republic of Interdisciplinary Program in Artificial Intelligence Seoul National University Seoul Korea Republic of

Transformer-based Super-Resolution (SR) methods have demonstrated superior performance compared to convolutional neural network (CNN)-based SR approaches due to their capability to capture long-range dependencies. However, their high computational complexity necessitates the development of lightweight approaches for practical use. To address this challenge, we propose the Attention-Sharing Information Distillation (ASID) network, a lightweight SR network that integrates attention-sharing and an information distillation structure specifically designed for Transformer-based SR methods. We modify the information distillation scheme, originally designed for efficient CNN operations, to reduce the computational load of stacked self-attention layers, effectively addressing the efficiency bottleneck. Additionally, we introduce attention-sharing across blocks to further minimize the computational cost of self-attention operations. By combining these strategies, ASID achieves competitive performance with existing SR methods while requiring only around 300 K parameters – significantly fewer than existing CNN-based and Transformer-based SR models. Furthermore, ASID outperforms state-of-the-art SR methods when the number of parameters is matched, demonstrating its efficiency and effectiveness. The code and supplementary material are available on the project page. Copyright © 2025, The Authors. All rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Frequency-Domain Multi-Exposure HDR Imaging Network With Representative Image Features

引用

IEEE ACCESS 2023年 11卷 124899-124910页

作者： Lee, Keuntek Park, Jaehyun Jang, Yeong Il Cho, Nam Ik Seoul Natl Univ Dept Elect & Comp Engn INMC Seoul 08826 South Korea Seoul Natl Univ Interdisciplinary Program Artificial Intelligence INMC Seoul 08826 South Korea

Constructing a high dynamic range (HDR) image from multi-exposure low dynamic range (LDR) images is challenging mainly due to two major problems. One is a large misalignment between the LDR images taken at different moments due to camera and object motions. The other is missing content in over- or under-exposed areas, which can be very large in the LDR images with long or short exposure times. These problems lead to ghosting artifacts or saturated regions in the reconstructed HDR image. In this paper, we propose a method using convolutional neural networks (CNNs) to address these problems. Specifically, each LDR image is fed to a two-branch CNN that extracts two kinds of features separately, named content features and global features. Then, another CNN aligns and merges the features to create the HDR output. The content features are used to create spatial attention masks, which are used to align the features in the time domain with exposure-invariant attributes. Further, Fourier coefficient vectors of global features are utilized to modulate intermediate features in the frequency domain during the reconstruction process. Experimental results show that our method achieves state-of-the-art HDR reconstruction performances on several benchmarks. Our code is available at https://***/KeuntekLee/FDM-HDR.

关键词： Feature extraction Image reconstruction Image processing Cameras Optical flow Brightness Frequency-domain analysis Convolutional neural networks frequency domain feature modulating high dynamic range imaging multi-exposure fusion

来源：评论

学校读者我要写书评

暂无评论

RFG-HDR: REPRESENTATIVE FEATURE-GUIDED TRANSFORMER FOR MULTI-EXPOSURE HIGH DYNAMIC RANGE IMAGING 31

RFG-HDR: REPRESENTATIVE FEATURE-GUIDED TRANSFORMER FOR MULTI...

引用

2024 International Conference on Image Processing

作者： Lee, Keuntek Park, Jaehyun Park, Gu Yong Cho, Nam Ik Seoul Natl Univ INMC Dept Elect & Comp Engn Seoul South Korea Seoul Natl Univ INMC Interdisciplinary Program Artificial Intelligence Seoul South Korea

ISBN: (纸本)9798350349405;9798350349399

Multi-exposure fusion is a high dynamic range (HDR) imaging technique that combines multiple low dynamic range (LDR) images of a scene with varying exposure times to produce a single high-quality HDR image. Since each LDR frame is captured with a different exposure time (bias), it is crucial to extract meaningful features from each differently-exposed LDR frame for producing high-quality HDR images. This paper introduces a new contrastive learning method that provides a versatile way of extracting characteristic features from LDR frames by considering the relationship between LDR frames. Additionally, we introduce Representative Feature-Guided Transformer (RFG-HDR), a new architecture that utilizes contrastive-learned representations to improve frame alignment and merging. Based on extensive experiments on various datasets, we have found that the RFG-HDR performs better than existing multi-exposure HDR imaging methods in terms of various evaluation metrics. Our work will be released on https://***/KeuntekLee/RFG-HDR.

关键词： HDR imaging multi-exposure fusion representation learning

来源：评论

学校读者我要写书评

暂无评论

Face Swapping for Low-Resolution and Occluded Images In-the-Wild

引用

IEEE ACCESS 2024年 12卷 91383-91395页

作者： Park, Jaehyun Kang, Wonjun Koo, Hyung Il Cho, Nam Ik Seoul Natl Univ Interdisciplinary Program Artificial Intelligence Seoul 08826 South Korea Seoul Natl Univ Dept Elect & Comp Engn INMC Seoul 08826 South Korea Ajou Univ Dept Elect & Comp Engn Suwon 16499 South Korea

Safeguarding personal identity in various surveillance videos, dashcams, and on-street videos is crucial. One way to do this is to detect faces and blur them, but a better solution is to replace them with non-existent ones to maintain the naturalness of the videos. While face swapping methods have already been used in the media industry with high-quality faces, it is challenging to apply them for identity protection to faces in-the-wild where faces are often occluded and of low-resolution. Therefore, we propose a new framework for face swapping specifically designed to work with face images taken in real-world scenarios, making it useful as a privacy protection method. To tackle the issue of low-resolution images, we introduce a Cross-Resolution Contrastive Loss (CRCL) technique, which allows our neural network model to be trained using triplets of varying resolutions. This enables the model to learn and use identity information across different resolutions, thereby improving its accuracy. We also propose a plug-and-play framework that can be easily applied to existing face swapping models to handle occlusions. By explicit swapping of facial features and filling of occluded regions, our framework provides a more seamless blend. To demonstrate the effectiveness of our method in handling faces in-the-wild, we create an occluded VGGFace2 dataset consisting of face images augmented with various facial masks and hand occlusions. Through quantitative and qualitative assessments on this dataset, our proposed method demonstrates robust performance under low-resolution or occluded scenarios. Significant improvements are made in the quality of swapped faces while preserving their identity and attributes, highlighting the effectiveness of our framework in advancing face swapping as a reliable privacy protection measure.

关键词： Image resolution Protection Privacy Feature extraction Videos Media Deep learning Identification of persons de-identification face swapping in-the-wild low-resolution occlusion privacy protection

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：