Qihe Liu

Robust Dynamic Expansion for Continual Learning under Backdoor Attacks via Purification and Selective Recovery

Keyu Lin, Fei Ye, Qihe Liu, Shijie Zhou et autres

Continual learning (CL) enables models to acquire new knowledge from sequentially arriving tasks while retaining previously learned knowledge. However, in practical scenarios, task streams collected from untrusted sources may contain backdoor-poisoned samples, posing a critical challenge to the stability, plasticity, and security …

Continual learning via semantic memory system

Mingsen Luo, Qihe Liu, Fei Ye, Adrian G. Borş et autres

cn, gb (code pays fourni par la source)

Continual learning via dynamic expandable task-specific and general representations

Qihe Liu, yongcheng zhong, Fei Ye, Adrian G. Borş et autres

cn, gb (code pays fourni par la source)

TempJail: Temporal Jailbreak Attack against Large Vision-Language Models via Subtitle Scheduling

Ling Zhou, Yihao Huang, Jingling Sun, Zhiwen Tian et autres

Large vision-language models (LVLMs) have achieved remarkable progress in video understanding and reasoning. Despite extensive studies on text- and image-based jailbreaks, video jailbreaks against LVLMs remain largely unexplored. Existing video jailbreak methods mainly manipulate textual content embedded in videos, while overlooking how …

Continual Learning across multiple domains via a Dynamic Expandable and Mergeable Model

Fei Ye, Ruilong Yu, Qihe Liu, Adrian Bors et autres

cn, gb, ae, us (code pays fourni par la source)

Synergistic Fourier-Wavelet Neural Operator

Tianle Lin, Ru Fang, Jiajun Sun, Qihe Liu et autres

Solving partial differential equations (PDEs) with neural operators has shown great promise, but existing methods face a trade-off. Fourier Neural Operators (FNOs) excel at capturing global, periodic features but struggle with localized, sharp gradients. Conversely, wavelet-based operators handle local details well but …

cn (code pays fourni par la source)

Learning Adaptive and Expandable Mixture Model for Continual Learning

Fei Ye, yongcheng zhong, Qihe Liu, Adrian G. Borş et autres

Continuous learning constitutes a fundamental capability of artificial intelligence systems, enabling them to incrementally assimilate novel information without succumbing to catastrophic forgetting. Recent research has leveraged Pre-Trained Models (PTMs) to enhance continual learning efficacy. Nevertheless, prevailing methodologies typically depend on a singular …

cn, gb (code pays fourni par la source)

Continual Learning Via Gradient-Regularized Based Dynamic Expansion Model

Fei Ye, yongcheng zhong, Qihe Liu, Adrian G. Borş et autres

cn, gb (code pays fourni par la source)

Improving Adversarial Robustness by Mitigating Instability Through Relearning

Yi Zeng, Ling Zhou, Ruilong Yu, Qihe Liu et autres

cn (code pays fourni par la source)

EPINN: Enhanced Physics-Informed Neural Network for Solving Continuous Integral Equations

Zhiyuan Ren, Shijie Zhou, Dong Liu, Qihe Liu

Background: Integral equations play a crucial role in modeling complex systems across various scientific disciplines. However, traditional numerical methods and existing physics-informed neural networks (PINNs) face substantial challenges, including the curse of dimensionality, uncontrolled error propagation, and limited generalization capabilities. Objectives: This …

cn (code pays fourni par la source)

A survey on closed-loop intelligent frameworks for parallel training of deep neural networks

Zhiyuan Ren, Shijie Zhou, Dong Liu, Qihe Liu

cn (code pays fourni par la source)

Physics-Informed Neural Networks: A Review of Methodological Evolution, Theoretical Foundations, and Interdisciplinary Frontiers Toward Next-Generation Scientific Computing

Zhiyuan Ren, Shijie Zhou, Dong Liu, Qihe Liu

Physics-informed neural networks (PINNs) have emerged as a transformative methodology integrating deep learning with scientific computing. This review establishes a three-dimensional analytical framework to systematically decode PINNs’ development through methodological innovation, theoretical breakthroughs, and cross-disciplinary convergence. The contributions include threefold: First, identifying …

cn (code pays fourni par la source)