In conjunction with the 4th Workshop on Visual Understanding by Learning from Web Data at CVPR2020.

Please find all details and links here .

The idea of this challenge is to bring together researcher in the fields of weak learning from image and video data, including but not limited to:

• Learning from instructional videos
• Learning from web data
• Learning from sound / spoken words
• Learning from weak labels
• Learning from noisy data
• Learning from natural labels

to explore their ideas and approaches on real world data.

To this end, we provide a dataset based on 20K real-live YouTube videos with various options to explore your systems on. This includes:

• Videos and pre-extracted features
• Audio data (together with videos) and text
• Test data of 250 full videos (hand annotated, 512 action classes)
• Sample code to get a baseline system running in 1-2h

We run two tracks in this challenge:

Video track. For the video track you are allowed to use any video/image data provided in the dataset. This can be the full videos or the video clips. You can train any CNN architecture and make use of the provided subtitles.

Feature track. For the feature track, you are only allowed to use the precomputed features as well as any subtitle information of the clips or the full videos.

For both tracks, you are allowed to use additional text data and knowledge sources that are publicly available. For details, please checkout the General rules and the FAQ section.

Given the current situation almost everywhere, we decided to match the price money with a charity donation of $500. Every winner can name an official charity organization of their choice that the money will be issued to.

Video track:

• 1st 1000 + 500 charity
• 2nd 500 + 500 charity

Feature track :

• 1st 1000 + 500 charity
• 2nd 500 + 500 charity

The tentative dates for the challenge are:

• Challenge Submissions Deadline: June 7, 2020
• Challenge Award Notification: June 10, 2020

All deadlines are at 23:59 Pacific Standard Time.

We will inform all participants through the website and codalab about any changes of the dates.

Submission to the poster paper track does NOT require participation in the challenge track.

The submission can be published or unpublished work, but have to be accessible publicly. We recommend authors to upload their paper on arXiv, but other publicly accessible link is also acceptable.

There is no requirement on paper format or page limitation. We recommend the CVPR formatting style with 4-8 pages.

The submission will be reviewed by workshop chairs. Accepted papers will be presented at the poster session at the workshop. Note that all accepted papers will be linked on the workshop website, but will NOT appear in the CVPR workshop proceedings.

Poster papers are reviewed in a rolling base until the places are fulfilled. Acceptance notification will be sent out once the decision has been made. We encourage people to submit as early as possible. For papers submitted before May 15th, 2020, the acceptance notification will be sent out at the latest by May 30th, 2020.

For details, please see here :

Poster Paper

• Paper Submission Deadline May 15, 2020
• Paper Acceptance Notification May 30, 2020

* All deadlines are at 23:59 Pacific Standard Time.

We look forward to your participation!

If you have any questions or comments, please contact kuehne@ibm.com