Label Denoising with Large Ensembles of Heterogeneous Neural Networks
This work addresses video understanding for applications like content labeling, but it is incremental as it builds on existing methods like ensembles and knowledge distillation.
The paper tackled the challenge of large-scale video classification on the YouTube-8M dataset by developing a top solution for a Kaggle competition, using a large ensemble of heterogeneous neural networks that fit within hardware constraints and achieved competitive results through techniques like knowledge distillation and mixup.
Despite recent advances in computer vision based on various convolutional architectures, video understanding remains an important challenge. In this work, we present and discuss a top solution for the large-scale video classification (labeling) problem introduced as a Kaggle competition based on the YouTube-8M dataset. We show and compare different approaches to preprocessing, data augmentation, model architectures, and model combination. Our final model is based on a large ensemble of video- and frame-level models but fits into rather limiting hardware constraints. We apply an approach based on knowledge distillation to deal with noisy labels in the original dataset and the recently developed mixup technique to improve the basic models.