Variational Stochastic Gradient Descent for Deep Neural Networks
This work addresses optimization challenges in deep learning, offering a potential improvement for training neural networks, though it appears incremental as it builds on existing adaptive and probabilistic approaches.
The paper tackled the problem of improving gradient-based optimization for deep neural networks by proposing Variational Stochastic Gradient Descent (VSGD), which combines probabilistic modeling with adaptive methods, and showed that VSGD outperforms Adam and SGD on image classification tasks with two datasets and four architectures.
Current state-of-the-art optimizers are adaptive gradient-based optimization methods such as Adam. Recently, there has been an increasing interest in formulating gradient-based optimizers in a probabilistic framework for better modeling the uncertainty of the gradients. Here, we propose to combine both approaches, resulting in the Variational Stochastic Gradient Descent (VSGD) optimizer. We model gradient updates as a probabilistic model and utilize stochastic variational inference (SVI) to derive an efficient and effective update rule. Further, we show how our VSGD method relates to other adaptive gradient-based optimizers like Adam. Lastly, we carry out experiments on two image classification datasets and four deep neural network architectures, where we show that VSGD outperforms Adam and SGD.