TAdam: A Robust Stochastic Gradient Optimizer
This work addresses robustness to noise for machine learning practitioners, but it is incremental as it modifies an existing optimizer.
The authors tackled the problem of noise in machine learning, particularly in robotics, by proposing TAdam, a robust stochastic gradient optimizer based on the student-t distribution, which outperformed Adam on regression, classification, and reinforcement learning tasks.
Machine learning algorithms aim to find patterns from observations, which may include some noise, especially in robotics domain. To perform well even with such noise, we expect them to be able to detect outliers and discard them when needed. We therefore propose a new stochastic gradient optimization method, whose robustness is directly built in the algorithm, using the robust student-t distribution as its core idea. Adam, the popular optimization method, is modified with our method and the resultant optimizer, so-called TAdam, is shown to effectively outperform Adam in terms of robustness against noise on diverse task, ranging from regression and classification to reinforcement learning problems. The implementation of our algorithm can be found at https://github.com/Mahoumaru/TAdam.git