Interpretable compositional convolutional neural networks
W Shen, Z Wei, S Huang, B Zhang, J Fan… - arXiv preprint arXiv …, 2021 - arxiv.org
The reasonable definition of semantic interpretability presents the core challenge in
explainable AI. This paper proposes a method to modify a traditional convolutional neural …
explainable AI. This paper proposes a method to modify a traditional convolutional neural …
Synchronization-Inspired Interpretable Neural Networks
Synchronization is a ubiquitous phenomenon in nature that enables the orderly presentation
of information. In the human brain, for instance, functional modules such as the visual, motor …
of information. In the human brain, for instance, functional modules such as the visual, motor …
Learning Task-preferred Inference Routes for Gradient De-conflict in Multi-output DNNs
Y Sun, X Xu, J Li, X Hu, Y Shi, LL Zeng - arXiv preprint arXiv:2305.19844, 2023 - arxiv.org
Multi-output deep neural networks (MONs) contain multiple task branches, and these tasks
usually share partial network filters that lead to the entanglement of different task inference …
usually share partial network filters that lead to the entanglement of different task inference …
Continual Learning through Networks Splitting and Merging with Dreaming-Meta-Weighted Model Fusion
Y Sun, X Xu, J Li, G Xie, Y Shi, Q Fang - arXiv preprint arXiv:2312.07082, 2023 - arxiv.org
It's challenging to balance the networks stability and plasticity in continual learning
scenarios, considering stability suffers from the update of model and plasticity benefits from …
scenarios, considering stability suffers from the update of model and plasticity benefits from …
[PDF][PDF] 深度自适应动态神经网络进展综述
孙毅, 李健, 徐昕, 王宇茹 - Laser & Optoelectronics Progress, 2022 - researching.cn
摘要深度神经网络极大地推动了视觉感知与自然语言处理任务的发展, 但当前大多数深度模型
执行的是静态推理图, 即网络结构(深度) 在推理阶段固定不变. 这种静态推理模式使模型无法 …
执行的是静态推理图, 即网络结构(深度) 在推理阶段固定不变. 这种静态推理模式使模型无法 …
Relative Contribution Mechanism: A Unified Paradigm for Disassembling Convolutional Neural Networks
With the tremendous development of CNNs, obtaining an available CNN classifier is more
challenging due to the massive number of parameters and deep structure. Recently, an …
challenging due to the massive number of parameters and deep structure. Recently, an …
Design and implementation of question recommendation system based on deep knowledge tracing
S Guo - 2022 3rd International Conference on Computer Vision …, 2022 - ieeexplore.ieee.org
Online education has developed rapidly since 2020, and the completion of after-school
exercises is a part of online education, which plays an important role in improving students' …
exercises is a part of online education, which plays an important role in improving students' …