Implicit Deep Learning
Implicit Deep Learning
复制标题
DOI：
10.1137/20m1358517
复制
发表时间：
2021-01-01
影响因子：
3.6
通讯作者：
Tsai, Alicia
中科院分区：
文献类型：
--
作者：
El Ghaoui, Laurent;Gu, Fangda;Tsai, Alicia
隐式深度学习预测规则概括了前馈神经网络的递归规则。这种规则是基于一个固定点方程的解决方案，涉及一个单一的隐藏特征向量，因此只有隐含定义。隐式框架大大简化了深度学习的符号，并在新架构和算法、鲁棒性分析和设计、可解释性、稀疏性和网络架构优化方面开辟了许多新的可能性。
Implicit deep learning prediction rules generalize the recursive rules of feedforward neural networks. Such rules are based on the solution of a fixed-point equation involving a single vector of hidden features, which is thus only implicitly defined. The implicit framework greatly simplifies the notation of deep learning, and opens up many new possibilities in terms of novel architectures and algorithms, robustness analysis and design, interpretability, sparsity, and network architecture optimization.