Zobrazeno 1 - 10
of 14 693
pro vyhledávání: '"Shien An"'
Deep learning models are known to be vulnerable to adversarial attacks by injecting sophisticated designed perturbations to input data. Training-time defenses still exhibit a significant performance gap between natural accuracy and robust accuracy. I
Externí odkaz:
http://arxiv.org/abs/2410.16805
Autor:
Li, Jonathan Weiping, Liang, Ren-Wei, Yeh, Cheng-Han, Tsai, Cheng-Chang, Yu, Kuanchun, Lu, Chun-Shien, Chen, Shang-Tse
This paper examines the phenomenon of probabilistic robustness overestimation in TRADES, a prominent adversarial training method. Our study reveals that TRADES sometimes yields disproportionately high PGD validation accuracy compared to the AutoAttac
Externí odkaz:
http://arxiv.org/abs/2410.07675
In recent years, Vision-Language Models (VLMs) have demonstrated significant advancements in artificial intelligence, transforming tasks across various domains. Despite their capabilities, these models are susceptible to jailbreak attacks, which can
Externí odkaz:
http://arxiv.org/abs/2410.01438
Autor:
Chen, Bang-Shien, Lin, Yu-Kai, Chen, Jian-Yu, Huang, Chih-Wei, Chern, Jann-Long, Sun, Ching-Cherng
Publikováno v:
IEEE Robotics and Automation Letters, 9(12), 11666-11673, 2024
Robust estimation is essential in computer vision, robotics, and navigation, aiming to minimize the impact of outlier measurements for improved accuracy. We present a fast algorithm for Geman-McClure robust estimation, FracGM, leveraging fractional p
Externí odkaz:
http://arxiv.org/abs/2409.13978
Visual State Space Model (VSS) has demonstrated remarkable performance in various computer vision tasks. However, in the process of development, backdoor attacks have brought severe challenges to security. Such attacks cause an infected model to pred
Externí odkaz:
http://arxiv.org/abs/2408.11679
Amid the proliferation of forged images, notably the tsunami of deepfake content, extensive research has been conducted on using artificial intelligence (AI) to identify forged content in the face of continuing advancements in counterfeiting technolo
Externí odkaz:
http://arxiv.org/abs/2407.18614
Deep learning technology has brought convenience and advanced developments but has become untrustworthy due to its sensitivity to adversarial attacks. Attackers may utilize this sensitivity to manipulate predictions. To defend against such attacks, e
Externí odkaz:
http://arxiv.org/abs/2407.15524
Semi-supervised learning (SSL) has achieved remarkable performance with a small fraction of labeled data by leveraging vast amounts of unlabeled data from the Internet. However, this large pool of untrusted data is extremely vulnerable to data poison
Externí odkaz:
http://arxiv.org/abs/2407.10180
Autor:
Fang, Yuwei, Menapace, Willi, Siarohin, Aliaksandr, Chen, Tsai-Shien, Wang, Kuan-Chien, Skorokhodov, Ivan, Neubig, Graham, Tulyakov, Sergey
Existing text-to-video diffusion models rely solely on text-only encoders for their pretraining. This limitation stems from the absence of large-scale multimodal prompt video datasets, resulting in a lack of visual grounding and restricting their ver
Externí odkaz:
http://arxiv.org/abs/2407.06304
Autor:
Chen, Tsai-Shien, Siarohin, Aliaksandr, Menapace, Willi, Deyneka, Ekaterina, Chao, Hsiang-wei, Jeon, Byung Eun, Fang, Yuwei, Lee, Hsin-Ying, Ren, Jian, Yang, Ming-Hsuan, Tulyakov, Sergey
The quality of the data and annotation upper-bounds the quality of a downstream model. While there exist large text corpora and image-text pairs, high-quality video-text data is much harder to collect. First of all, manual labeling is more time-consu
Externí odkaz:
http://arxiv.org/abs/2402.19479