|
Briefing
|
- Deepfake Detection Challenge (DFDC) – Facebook and partners Microsoft, the Partnership on AI, and several universities joined together to setup contest for developing technologies that can spot fake artificial intelligence (AI) generated videos
- Partnering Institutions – Include Cornell Tech, MIT, University of Oxford, UC Berkeley, University of Maryland, College Park, and University at Albany-SUNY
- Building Datasets – Aims to create datasets of deep fake videos using paid actors to test technologies, but will not leverage user data from Facebook
- Rewards – Will publish leaderboard featuring top technologies, with winners getting grants and awards
- $10 Million Investment – Facebook is contributing over $10 million in investment to fund the challenge
- Launch Schedule – Partial dataset and challenge parameters will initially be tested at International Conference on Computer Vision (ICCV) in October 2019, with full datasets released and complete launch to occur at Conference on Neural Information Processing Systems (NeurIPS) in December 2019
|
|
Accelerator
|
|
|
Business Model and Practices
Business Model
and Practices
|
|
|
Sector
|
Information Technology
|
|
Function
|
Research and Development
|
|
Organization
|
Cornell Tech, Facebook Inc., Massachusetts Institute of Technology, Microsoft Corp., Partnership on AI, University at Albany - State University of New York (SUNY), University of California - Berkeley, University of Maryland, University of Oxford
|
|
Source
|
-
Schroepfer, M., "Creating a data set and a challenge for deepfakes",
-
Culliford, E., "Facebook, Microsoft launch contest to detect deepfake videos",
-
Coldewey, D., "Facebook is making its own deepfakes and offering prizes for detecting them",
-
AcceleratingBiz analysis
|
|
Original Publication Date
|
September 5, 2019
|
