Hang Su

Hang Su

Auto-generated profile, based on publicly available information

Hang Su

Tsinghua University

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
26 February 2024 in IEEE Transactions on Pattern Analysis and Machine Intelligence
Verified
Journal Article
Closed access
LWLiyuan Wang
Liyuan Wang
XZXingxing Zhang
Xingxing Zhang
HSHang Su
Hang Su
JZJun Zhu
Jun Zhu
To cope with real-world dynamics, an intelligent system needs to incrementally acquire, update, accumulate, and exploit knowledge throughout its lifetime. This ability, known as continual learning, provides a foundation for AI systems to develop themselves adaptively. In a general sense, continual learning is explicitly limited by catastrophic forgetting, where learning a new task usually results in a dramatic performance drop of the old tasks. Beyond this, increasingly numerous advances have emerged in recent years that largely extend the understanding and application of continual learning. The growing and widespread interest in this direction demonstrates its realistic significance as well as complexity. In this work, we present a comprehensive survey of continual learning, seeking to bridge the basic settings, theoretical foundations, representative methods, and practical applications. Based on existing theoretical and empirical results, we summarize the general objectives of continual learning as ensuring a proper stability-plasticity trade-off and an adequate intra/inter-task generalizability in the context of resource efficiency. Then we provide a state-of-the-art and elaborated taxonomy, extensively analyzing how representative strategies address continual learning, and how they are adapted to particular challenges in various applications. Through an in-depth discussion of promising directions, we believe that such a holistic perspective can greatly facilitate subsequent exploration in this field and beyond.

9 August 2024 in International Journal of Computer Vision
Verified
Journal Article
Closed access
PDF file
CLChang Liu
Chang Liu
YDYinpeng Dong
Yinpeng Dong
WXWenzhao Xiang
Wenzhao Xiang
XYXiao Yang
Xiao Yang
HSHang Su
Hang Su
SZShibao Zheng
Shibao Zheng

26 October 2023
Verified
Conference Paper
Open access
SZShengfang Zhai
Shengfang Zhai
YDYinpeng Dong
Yinpeng Dong
QSQingni Shen
Qingni Shen
SPShi Pu
Shi Pu
YFYuejian Fang
Yuejian Fang
HSHang Su
Hang Su
With the help of conditioning mechanisms, the state-of-the-art diffusion models have achieved tremendous success in guided image generation, particularly in text-to-image synthesis. To gain a better understanding of the training process and potential risks of text-to-image synthesis, we perform a systematic investigation of backdoor attack on text-to-image diffusion models and propose BadT2I, a general multimodal backdoor attack framework that tampers with image synthesis in diverse semantic levels. Specifically, we perform backdoor attacks on three levels of the vision semantics: Pixel-Backdoor, Object-Backdoor and Style-Backdoor. By utilizing a regularization loss, our methods efficiently inject backdoors into a large-scale text-to-image diffusion model while preserving its utility with benign inputs. We conduct empirical experiments on Stable Diffusion, the widely-used text-to-image diffusion model, demonstrating that the large-scale diffusion model can be easily backdoored within a few fine-tuning steps. We conduct additional experiments to explore the impact of different types of textual triggers, as well as the backdoor persistence during further training, providing insights for the development of backdoor defense methods. Besides, our investigation may contribute to the copyright protection of text-to-image models in the future. Our Code: https://github.com/sf-zhai/BadT2I.

23 January 2020 in IEEE Transactions on Visualization and Computer Graphics
Verified
Journal Article
Closed access
KCKelei Cao
Kelei Cao
MLMengchen Liu
Mengchen Liu
HSHang Su
Hang Su
JWJing Wu
Jing Wu
JZJun Zhu
Jun Zhu
SLShixia Liu
Shixia Liu
Adversarial examples, generated by adding small but intentionally imperceptible perturbations to normal examples, can mislead deep neural networks (DNNs) to make incorrect predictions. Although much work has been done on both adversarial attack and defense, a fine-grained understanding of adversarial examples is still lacking. To address this issue, we present a visual analysis method to explain why adversarial examples are misclassified. The key is to compare and analyze the datapaths of both the adversarial and normal examples. A datapath is a group of critical neurons along with their connections. We formulate the datapath extraction as a subset selection problem and solve it by constructing and training a neural network. A multi-level visualization consisting of a network-level visualization of data flows, a layer-level visualization of feature maps, and a neuron-level visualization of learned features, has been designed to help investigate how datapaths of adversarial and normal examples diverge and merge in the prediction process. A quantitative evaluation and a case study were conducted to demonstrate the promise of our method to explain the misclassification of adversarial examples.

1 October 2023
Verified
Conference Paper
Open access
SLShilong Liu
Shilong Liu
TRTianhe Ren
Tianhe Ren
JCJiayu Chen
Jiayu Chen
ZZZhaoyang Zeng
Zhaoyang Zeng
HZHao Zhang
Hao Zhang
LZLei Zhang
Lei Zhang
This paper is concerned with the matching stability problem across different decoder layers in DEtection TRansformers (DETR). We point out that the unstable matching in DETR is caused by a multi-optimization path problem, which is highlighted by the one-to-one matching design in DETR. To address this problem, we show that the most important design is to use and only use positional metrics (like IOU) to supervise classification scores of positive examples. Under the principle, we propose two simple yet effective modifications by integrating positional metrics to DETR’s classification loss and matching cost, named position-supervised loss and position-modulated cost. We verify our methods on several DETR variants. Our methods show consistent improvements over baselines. By integrating our methods with DINO, we achieve 50.4 and 51.5 AP on the COCO detection benchmark using ResNet-50 backbones under 1× (12 epochs) and 2× (24 epochs) training settings, achieving a new record under the same setting. We achieve 63.8 AP on COCO detection test-dev with a Swin-Large backbone. Our code will be made available at https://github.com/IDEA-Research/Stable-DINO.

20 May 2019 in IEEE Transactions on Signal Processing
Verified
Journal Article
Closed access
YZYi Zhou
Yi Zhou
TWTian Wang
Tian Wang
RHRonghua Hu
Ronghua Hu
HSHang Su
Hang Su
YLYi Liu
Yi Liu
HSHichem Snoussi
Hichem Snoussi
Conventional tracking filter in marine radar only concerns the position or shape of the object. When the target encounters the clutter interruption or ship occlusion, the tracker is easy to drift away due to the wrong target association. Since in short-range maritime surveillance, X-band marine radar captures the object in extended region with varying intensities, in this paper, combining position, shape, and appearance of the target together, multiple kernelized correlation filters (MKCF) are proposed to conduct a single object tracking in the real marine radar. By automatic initializing KCF on a target in different time steps and fusing these multiple KCFs via the maximum likelihood, the proposed tracker implicitly uses multiple instances of the target to improve the robustness of the long-term tracking. Bounding rectangles of the multiple trackers also enhance the reliability of ship segmentation via a voting procedure. In the real ship tracking experiment, the proposed tracker performs favorably against the conventional radar tracker and top-ranked visual tracker in the cases of clutter interruption, ship occlusions, and scale changing. We believe that our proposal sheds light on that the intensity distribution of the extended object could be valuable for target association in marine radar data. To encourage the further researching, our tracking framework is made open-source.