Can you produce the best video tag predictions?
Video captures a cross-section of our society. And major advances in analyzing and understanding video have the potential to touch all aspects of life from learning and communication to entertainment and play. In this competition, Google is inviting the Kaggle community to join efforts to accelerate research in large-scale video understanding, while giving participants early access to the Google Cloud Machine Learning (Cloud ML) beta platform.
Today, one of the greatest obstacles to rapid improvements in video understanding research has been the lack of large-scale, labeled datasets open to the public. For example, the availability of large, labeled datasets such as ImageNet has enabled continued breakthroughs in machine learning and machine perception. To that end, Google’s recent release of the YouTube-8M (YT-8M) dataset represents a significant step in this direction. Making this resource open to everyone from students and industry professionals is expected to kickstart innovation in areas such as representation learning and video modeling architectures.
In this competition, you are challenged to develop classification algorithms which accurately assign video-level labels using the new and improved YT-8M V2 dataset. The dataset was created from over 7 million YouTube videos (450,000 hours of video) and includes video labels from a vocabulary of 4716 classes (3.4 labels/video on average). It also comes with pre-extracted audio & visual features from every second of video (3.2B feature vectors in total). By taking part, Kagglers will not only play a pivotal role in setting state-of-the-art benchmarks, but also improve search and organization of video archives.
for more Details :