Frequently Asked Questions
What is the goal of the Deepfake Detection Challenge?
The AI technologies that power deepfakes and other tampered media are rapidly evolving, making deepfakes so hard to detect that, at times, even human evaluators can’t reliably tell the difference. The Deepfake Detection Challenge is designed to incentivize rapid progress in this area by inviting participants to compete to create new ways of detecting and preventing manipulated media.
When will the challenge begin?
We announced the release of the full dataset and the competition on December 11th. Please see the Kaggle competition page for a timeline of key dates and details on how to participate.
When is the submission deadline?
The challenge will run through the end of March 2020. Please see the Kaggle competition page for a timeline of key dates and details on how to participate.
Who can participate in the challenge?
The challenge is global and participants need to have a Kaggle account to access the dataset and participate in the challenge. Participants will also need to agree to our dataset license before participating in the challenge.
How do I join the challenge?
The challenge is available on Kaggle. To join the challenge, you must have a Kaggle account. You cannot use the credentials used for the preview dataset to participate in the competition and get access to the full dataset.
How was the training dataset created?
We constructed a new training dataset specifically for this challenge. To create this dataset, we worked with a third-party vendor to engage a diverse set of individuals who have agreed to participate in creating the dataset for this challenge. We then created tampered videos based on a subset of these unmodified videos, using a variety of different AI techniques.
Are you using user data from social media or video platforms in the dataset?
No user data from social or video platforms was included in the training dataset. We constructed a new dataset specifically for this challenge.
How will the challenge be judged and a winner selected?
Participants in the challenge must submit their code into a black box environment for testing. Please see the Kaggle competition page for more details on evaluation metrics and how winners will be selected.
What rights do challenge participants have to the tech they create for the challenge?
"Participants will retain rights to their models trained on the training dataset. Facebook and its subcontractors will receive rights from Participants to use the models to administer the challenge. For further details, please see the Kaggle competition rules.
How are you protecting against adversaries who will try to access the code and data?
Can I use the DFDC dataset for something other than this challenge?
The datasets can be used to research, develop and improve software, algorithms, machine learning models, techniques and technologies designed to detect manipulated media, images, audio and videos.
I received access to the preview dataset, how do I get access to the full dataset?
Participants need to have a Kaggle account to access the full dataset and participate in the challenge. You cannot use the credentials used for the preview dataset to participate in the competition and get access to the full dataset. Please see the Kaggle competition page to get access.
May licensees of the DFDC Dataset perform minor pre-processing or augmentations on the DFDC Dataset (e.g., blurring, rotations and color enhancements/saturations)?
Licensees of the DFDC Dataset may perform minor pre-processing or augmentations of the DFDC Dataset (or portions thereof); provided, that such minor pre-processing or augmentations are not for the purpose of creating a new dataset (or portion thereof) and are only used for the authorized “Purpose” stated within the Deepfake Detection Challenge Dataset License Agreement. Licensees of the DFDC Dataset may also modify the DFDC Dataset solely for the purpose of enabling compatibility with Licensees’ computer systems