Large-Scale Video Classification with Feature Space Augmentation coupled with Learned Label Relations and Ensembling
This work addresses video understanding for AI competitions, but it is incremental as it builds on existing methods with optimizations like ensembling and regularization.
The paper tackled large-scale video classification by developing a solution for the YouTube-8M challenge, achieving a global average precision of 88.733% on the private test set and ranking 3rd among 394 teams.
This paper presents the Axon AI's solution to the 2nd YouTube-8M Video Understanding Challenge, achieving the final global average precision (GAP) of 88.733% on the private test set (ranked 3rd among 394 teams, not considering the model size constraint), and 87.287% using a model that meets size requirement. Two sets of 7 individual models belonging to 3 different families were trained separately. Then, the inference results on a training data were aggregated from these multiple models and fed to train a compact model that meets the model size requirement. In order to further improve performance we explored and employed data over/sub-sampling in feature space, an additional regularization term during training exploiting label relationship, and learned weights for ensembling different individual models.