Song Gao

Song Gao

Auto-generated profile, based on publicly available information

Song Gao

Yunnan University

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

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

Representative Publications

View All Publications
16 February 2023 in Applied Intelligence
Verified
Journal Article
Closed access
PDF file
YZYi Zhao
Yi Zhao
Xin JinXin Jin
Xin Jin
SGSong Gao
Song Gao
LWliwen wu
liwen wu
SYShaowen Yao
Shaowen Yao
QJQian Jiang
Qian Jiang

1 February 2023 in IEEE Transactions on Dependable and Secure Computing
Verified
Journal Article
Open access
SGSong Gao
Song Gao
Ruxin WangRuxin Wang
Ruxin Wang
XWXiaoxuan Wang
Xiaoxuan Wang
SYShui Yu
Shui Yu
YDYunyun Dong
Yunyun Dong
WZWei Zhou
Wei Zhou
Despite achieving exceptional performance, deep neural networks (DNNs) suffer from the harassment caused by adversarial examples, which are produced by corrupting clean examples with tiny perturbations. Many powerful defense methods have been presented such as training data augmentation and input reconstruction which, however, usually rely on the prior knowledge of the targeted models or attacks. A clean example and its adversarial version are very similar but have different high-level representations in a victim model. If we can obtain a space in which the representations of similar examples are also similar, then adversarial examples can be picked out by comparing the representations of input examples in this space and the high-level space of the victim model. Inspired by this, we propose a novel approach for detecting adversarial images, which can protect any pre-trained DNN classifiers and resist an endless stream of new attacks. Specifically, we first adopt a dual autoencoder to project images to a latent space. The dual autoencoder uses the self-supervised learning to ensure that small modifications to samples do not significantly alter their latent representations. Next, the mutual information neural estimation is utilized to enhance the discrimination of the latent representations. We then leverage the prior distribution matching to regularize the latent representations. To easily compare the representations of examples in the two spaces, and not rely on the prior knowledge of the targeted model, a simple fully connected neural network is used to embed the learned representations into an eigenspace, which is consistent with the output eigenspace of the targeted model. Through the distribution similarity of an input example in the two eigenspaces, we can judge whether the input example is adversarial or not. Extensive experiments on MNIST, CIFAR-10, and ImageNet show that the proposed method has superior defense performance and transferability than state-of-the-arts.

30 November 2022 in Mathematics
Verified
Journal Article
Open access
PDF file
MZMa Zhiyu
Ma Zhiyu
SYShaowen Yao
Shaowen Yao
LWliwen wu
liwen wu
SGSong Gao
Song Gao
YZYUNQI ZHANG
YUNQI ZHANG
With the popularity of posting memes on social platforms, the severe negative impact of hateful memes is growing. As existing detection models have lower detection accuracy than humans, hateful memes detection is still a challenge to statistical learning and artificial intelligence. This paper proposed a multi-task learning method consisting of a primary multimodal task and two unimodal auxiliary tasks to address this issue. We introduced a self-supervised generation strategy in auxiliary tasks to generate unimodal auxiliary labels automatically. Meanwhile, we used BERT and RESNET as the backbone for text and image classification, respectively, and then fusion them with a late fusion method. In the training phase, the backward guidance technique and the adaptive weight adjustment strategy were used to capture the consistency and variability between different modalities, numerically improving the hateful memes detection accuracy and the generalization and robustness of the model. The experiment conducted on the Facebook AI multimodal hateful memes dataset shows that the prediction accuracy of our model outperformed the comparing models.

2 February 2023 in Entropy
Verified
Journal Article
Open access
PDF file
JYJingzong Yang
Jingzong Yang
CZChengjiang Zhou
Chengjiang Zhou
XLXueFeng Li
XueFeng Li
APAnning Pan
Anning Pan
TYTianqing Yang
Tianqing Yang
In modern industry, due to the poor working environment and the complex working conditions of mechanical equipment, the characteristics of the impact signals caused by faults are often submerged in strong background signals and noises. Therefore, it is difficult to effectivelyextract the fault features. In this paper, a fault feature extraction method based on improved VMD multi-scale dispersion entropy and TVD-CYCBD is proposed. First, the marine predator algorithm (MPA) is used to optimize the modal components and penalty factors in VMD. Second, the optimized VMD is used to model and decompose the fault signal, and then the optimal signal components are filtered according to the combined weight index criteria. Third, TVD is used to denoise the optimal signal components. Finally, CYCBD filters the de-noised signal and then envelope demodulation analysis is carried out. Through the simulation signal experiment and the actual fault signal experiment, the results verified that multiple frequency doubling peaks can be seen from the envelope spectrum, and there is little interference near the peak, which shows the good performance of the method.

8 October 2021 in IET Image Processing
Verified
Journal Article
Open access
PDF file
SGSong Gao
Song Gao
SYShui Yu
Shui Yu
LWLiwen Wu
Liwen Wu
SYShaowen Yao
Shaowen Yao
XZXiaowei Zhou
Xiaowei Zhou
Deep neural networks are widely adopted powerful tools for perceptual tasks. However, recent research indicated that they are easily fooled by adversarial examples, which are produced by adding imperceptible adversarial perturbations to clean examples. Here the steganalysis rich model (SRM) is utilized to generate noise feature maps, and they are combined with RGB images to discover the difference between adversarial examples and clean examples. In particular, a two‐stream pseudo‐siamese network that fuses the subtle difference in RGB images with the noise inconsistency in noise features is proposed. The proposed method has strong detection capability and transferability, and can be combined with any model without modifying its architecture or training procedure. The extensive empirical experiments show that, compared with the state‐of‐the‐art detection methods, the proposed approach achieves excellent performance in distinguishing adversarial samples generated by popular attack methods on different real datasets. Moreover, this method has good generalization, it trained by a specific adversary can defend against other adversaries effectively.

13 February 2024 in The Visual Computer
Verified
Journal Article
Closed access
PDF file
JJJun Ji
Jun Ji
SGSong Gao
Song Gao
WZWei Zhou
Wei Zhou