Carpe Diem, Seize the Samples Uncertain "at the Moment" for Adaptive Batch Selection

Hwanjun Song, Minseok Kim, Sundong Kim, Jae-Gil Lee

Keywords: batch selection, uncertain sample, acceleration, convergence

TL;DR: We explore the issue of truly uncertain samples for more effective batch selection.

Abstract: The performance of deep neural networks is significantly affected by how well mini-batches are constructed. In this paper, we propose a novel adaptive batch selection algorithm called Recency Bias that exploits the uncertain samples predicted inconsistently in recent iterations. The historical label predictions of each sample are used to evaluate its predictive uncertainty within a sliding window. By taking advantage of this design, Recency Bias not only accelerates the training step but also achieves a more accurate network. We demonstrate the superiority of Recency Bias by extensive evaluation on two independent tasks. Compared with existing batch selection methods, the results showed that Recency Bias reduced the test error by up to 20.5% in a fixed wall-clock training time. At the same time, it improved the training time by up to 59.3% to reach the same test error.