Our client, a leading social media company, wanted to test the fairness and robustness of their machine learning model by creating an open-source dataset. They partnered with us to collect thousands of audio and video samples from a diverse group of participants.
How We Do It
- Crowd Recruitment: We came up with a multi-faceted approach by creating personalized messaging to recruit 13 diverse groups of participants from India, the Philippines, and Vietnam through various marketing efforts.
- Crowdsourced Data Collection: We conducted in-person events in India and the Philippines to attract more participants.
- Data Quality: We ran manual and automated checks against all submitted data.
- Data Management & Security: We ensured all captured data were handled according to appropriate privacy and security guidelines.
- 13 Different demographic groups covered
- 86.8% Participant Approval Rate
- 11,056 Videos approved and delivered to the client
- 11,800+ TaskVerse sign-ups through organic and marketing efforts
Download our case study, Crowdsourced Audio, and Video Data Collection for a Leading Social Media Company, to learn more about the custom operations framework we implemented for our client’s machine learning models.
Download Case Study