Video is one of the richest sources of information available online but extracting deep insights from video content at internet scale is still an open problem, both in terms of depth and breadth of understanding, as well as scale. Over the last few years, the field of video understanding has made great strides due to the availability of large-scale video datasets and core advances in image, audio, and video modeling architectures. However, the state-of-the-art architectures on small scale datasets are frequently impractical to deploy at internet scale, both in terms of the ability to train such deep networks on hundreds of millions of videos, and to deploy them for inference on billions of videos. In this paper, we present a MapReduce-based training framework, which exploits both data parallelism and model parallelism to scale training of complex video models. The proposed framework uses alternating optimization and full-batch fine-tuning, and supports large Mixture-of-Experts classifiers with hundreds of thousands of mixtures, which enables a trade-off between model depth and breadth, and the ability to shift model capacity between shared (generalization) layers and per-class (specialization) layers. We demonstrate that the proposed framework is able to reach state-of-the-art performance on the largest public video datasets, YouTube-8M and Sports-1M, and can scale to 100 times larger datasets.
|Title of host publication||KDD 2019 - Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining|
|Publisher||Association for Computing Machinery|
|Number of pages||9|
|Publication status||Published - 2019 Jul 25|
|Event||25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2019 - Anchorage, United States|
Duration: 2019 Aug 4 → 2019 Aug 8
|Name||Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining|
|Conference||25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2019|
|Period||19/8/4 → 19/8/8|
Bibliographical notePublisher Copyright:
© 2019 Copyright held by the owner/author(s).
All Science Journal Classification (ASJC) codes
- Information Systems