The robustness of deep neural networks is usually lacking under adversarial examples, common corruptions, and distribution shifts, which becomes an important research problem in the development of deep learning. Although new deep learning methods and robustness improvement techniques have been constantly proposed, the robustness evaluations of existing methods are often inadequate due to their rapid development, diverse noise patterns, and simple evaluation metrics. Without thorough robustness evaluations, it is hard to understand the advances in the field and identify the effective methods. In this paper, we establish a comprehensive robustness benchmark called \textbf{ARES-Bench} on the image classification task. In our benchmark, we evaluate the robustness of 55 typical deep learning models on ImageNet with diverse architectures (e.g., CNNs, Transformers) and learning algorithms (e.g., normal supervised training, pre-training, adversarial training) under numerous adversarial attacks and out-of-distribution (OOD) datasets. Using robustness curves as the major evaluation criteria, we conduct large-scale experiments and draw several important findings, including: 1) there is an inherent trade-off between adversarial and natural robustness for the same model architecture; 2) adversarial training effectively improves adversarial robustness, especially when performed on Transformer architectures; 3) pre-training significantly improves natural robustness based on more training data or self-supervised learning. Based on ARES-Bench, we further analyze the training tricks in large-scale adversarial training on ImageNet. By designing the training settings accordingly, we achieve the new state-of-the-art adversarial robustness. We have made the benchmarking results and code platform publicly available.
GeoChrono: Benchmarking and Rethinking Long-Term Temporal Understanding in Remote Sensing
Beijing University of Posts and Telecommunications（北京邮电大学）；Tsinghua University（清华大学）；Hunan Normal University（湖南师范大学）；City University of Hong Kong（香港城市大学）
Trained on large datasets, deep learning (DL) can accurately classify videos into hundreds of diverse classes. However, video data is expensive to annotate. Zero-shot learning (ZSL) proposes one solution to this problem. ZSL trains a model once, and generalizes to new tasks whose classes are not present in the training dataset. We propose the first end-to-end algorithm for ZSL in video classification. Our training procedure builds on insights from recent video classification literature and uses a trainable 3D CNN to learn the visual features. This is in contrast to previous video ZSL methods, …
Rethinking Unsupervised Cross-modal Flow Estimation: Learning from Decoupled Optimization and Consistency Constraint
College of Information Science and Electronic Engineering, Zhejiang University（浙江大学信息科学与电子工程学院）；Ningbo Global Innovation Center, Zhejiang University（浙江大学宁波全球创新中心）；NingboTech University（宁波科技大学）
State Key Lab of Software Development Environment, Beihang University（北京航空航天大学软件开发环境国家重点实验室）；SenseTime（商汤科技）；Berkeley Artificial Intelligence Research Lab, University of California, Berkeley（加州大学伯克利分校伯克利人工智能研究实验室）；Johns Hopkins University（约翰斯·霍普金斯大学）；University of Oxford（牛津大学）；JD Explore Academy（京东探索研究院）
Deep neural networks (DNNs) are vulnerable to adversarial noises, which motivates the benchmark of model robustness. Existing benchmarks mainly focus on evaluating defenses, but there are no comprehensive studies of how architecture design and training techniques affect robustness. Comprehensively benchmarking their relationships is beneficial for better understanding and developing robust DNNs. Thus, we propose RobustART, the first comprehensive Robustness investigation benchmark on ImageNet regarding ARchitecture design (49 human-designed off-the-shelf architectures and 1200+ networks from n…
National University of Defense Technology（国防科技大学）；University of Oulu（奥卢大学）；Nanyang Technological University（南洋理工大学）；Stanford University（斯坦福大学）
Few-Shot Class-Incremental Learning (FSCIL) aims at incrementally learning novel classes from a few labeled samples by avoiding the overfitting and catastrophic forgetting simultaneously. The current protocol of FSCIL is built by mimicking the general class-incremental learning setting, while it is not totally appropriate due to the different data configuration, i.e., novel classes are all in the limited data regime. In this paper, we rethink the configuration of FSCIL with the open-set hypothesis by reserving the possibility in the first session for incoming categories. To assign better perfo…
Relative-Absolute Fusion: Rethinking Feature Extraction in Image-Based Iterative Method Selection for Solving Sparse Linear Systems
School of Software Technology, Dalian University of Technology（大连理工大学软件学院）；Greater Bay Area National Center of Technology Innovation（粤港澳大湾区国家技术创新中心）；CATARC Automotive Test Center (Tianjin) Co., Ltd（CATARC汽车测试中心（天津）有限公司）
Rethinking Intracranial Aneurysm Vessel Segmentation: A Perspective from Computational Fluid Dynamics Applications
Artificial Intelligence Innovation and Incubation Institute, Fudan University（复旦大学人工智能创新与孵化院）；Shanghai Academy of Artificial Intelligence for Science（上海人工智能科学研究院）；Department of Civil and Environmental Engineering, Hong Kong Polytechnic University（香港理工大学土木与环境工程系）；Huashan Hospital, Fudan University（复旦大学华山医院）
SenseTime Research（商汤科技研究院）；The University of Hong Kong（香港大学）；Sun Yat-Sen University（中山大学）；The Chinese University of Hong Kong（香港中文大学）
Channel pruning is a popular technique for compressing convolutional neural networks (CNNs), where various pruning criteria have been proposed to remove the redundant filters. From our comprehensive experiments, we found two blind spots in the study of pruning criteria: (1) Similarity: There are some strong similarities among several primary pruning criteria that are widely cited and compared. According to these criteria, the ranks of filters'Importance Score are almost identical, resulting in similar pruned structures. (2) Applicability: The filters'Importance Score measured by some pruning c…
Inception Institute of Artificial Intelligence（初始人工智能研究院）；University of Macau（澳门大学）；Terminus Group（特斯联集团）
The pioneering method for unsupervised meta-learning, CACTUs, is a clustering-based approach with pseudo-labeling. This approach is model-agnostic and can be combined with supervised algorithms to learn from unlabeled data. However, it often suffers from label inconsistency or limited diversity, which leads to poor performance. In this work, we prove that the core reason for this is lack of a clustering-friendly property in the embedding space. We address this by minimizing the inter- to intra-class similarity ratio to provide clustering-friendly embedding features, and validate our approach t…
LabelBench: A Comprehensive Framework for Benchmarking Adaptive Label-Efficient Learning
University of Wisconsin, Madison（威斯康星大学麦迪逊分校）；University of Washington（华盛顿大学）；Georgia Institute of Technology（佐治亚理工学院）；University of California, Riverside（加州大学河滨分校）
Rethinking Uncertainly Missing and Ambiguous Visual Modality in Multi-Modal Entity Alignment
Zhejiang University（浙江大学）；The University of Manchester（曼彻斯特大学）；University of Oxford（牛津大学）；The University of Edinburgh（爱丁堡大学）；Tsinghua University（清华大学）；Donghai laboratory（东海实验室）
Data augmentation is an effective way to improve the performance of deep networks. Unfortunately, current methods are mostly developed for high-level vision tasks (e.g., classification) and few are studied for low-level vision tasks (e.g., image restoration). In this paper, we provide a comprehensive analysis of the existing augmentation methods applied to the super-resolution task. We find that the methods discarding or manipulating the pixels or features too much hamper the image restoration, where the spatial relationship is very important. Based on our analyses, we propose CutBlur that cut…
In this paper, we examine the problem of real-world image deblurring and take into account two key factors for improving the performance of the deep image deblurring model, namely, training data synthesis and network architecture design. Deblurring models trained on existing synthetic datasets perform poorly on real blurry images due to domain shift. To reduce the domain gap between synthetic and real domains, we propose a novel realistic blur synthesis pipeline to simulate the camera imaging process. As a result of our proposed synthesis method, existing deblurring models could be made more r…
Zhejiang University（浙江大学）；Columbia University（哥伦比亚大学）；Hong Kong University of Science and Technology（香港科技大学）
Distinctive Image Captioning (DIC) -- generating distinctive captions that describe the unique details of a target image -- has received considerable attention over the last few years. A recent DIC work proposes to generate distinctive captions by comparing the target image with a set of semantic-similar reference images, i.e., reference-based DIC (Ref-DIC). It aims to make the generated captions can tell apart the target and reference images. Unfortunately, reference images used by existing Ref-DIC works are easy to distinguish: these reference images only resemble the target image at scene-l…
Benchmarking Large Vision-Language Models on Fine-Grained Image Tasks: A Comprehensive Evaluation
School of Computer Science and Engineering, School of Intelligence Science and Engineering and Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications, Southeast University, China（东南大学计算机科学与工程学院、智能科学与工程学院及新一代人工智能技术及其跨学科应用关键实验室，中国）；Wangxuan Institute of Computer Technology, Peking University, China（北京大学王轩计算机技术研究所，中国）；University of Copenhagen, Denmark（丹麦哥本哈根大学）
State-of-the-art face recognition methods typically take the multi-classification pipeline and adopt the softmax-based loss for optimization. Although these methods have achieved great success, the softmax-based loss has its limitation from the perspective of open set classification: the multi-classification objective in the training phase does not strictly match the objective of open set classification testing. In this paper, we derive a new loss named global boundary CosFace (GB-CosFace). Our GB-CosFace introduces an adaptive global boundary to determine whether two face samples belong to th…
3D perception, especially point cloud classification, has achieved substantial progress. However, in real-world deployment, point cloud corruptions are inevitable due to the scene complexity, sensor inaccuracy, and processing imprecision. In this work, we aim to rigorously benchmark and analyze point cloud classification under corruptions. To conduct a systematic investigation, we first provide a taxonomy of common 3D corruptions and identify the atomic corruptions. Then, we perform a comprehensive evaluation on a wide range of representative point cloud models to understand their robustness a…
MIBench: A Comprehensive Framework for Benchmarking Model Inversion Attack and Defense
School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen（哈尔滨工业大学（深圳）计算机科学与技术学院）；Tsinghua Shenzhen International Graduate School, Tsinghua University（清华大学深圳国际研究生院）；Department of Computer Science and Technology, Tsinghua University（清华大学计算机科学与技术系）
Quality Assured: Rethinking Annotation Strategies in Imaging AI
German Cancer Research Center（德国癌症研究中心）；Helmholtz Imaging（亥姆霍兹成像中心）；Heidelberg University（海德堡大学）；University of Minnesota（明尼苏达大学）；National Center for Tumor Diseases（国家肿瘤疾病中心）
MechVQA: Benchmarking and Enhancing Multimodal LLMs on Comprehensive Mechanical Drawing Understanding
Beijing Academy of Artificial Intelligence (BAAI), China（北京人工智能研究院）；Institute of Information Engineering, Chinese Academy of Sciences, China（信息工程研究所）；Beijing University of Technology, China（北京理工大学）
