-
Trove
an app for Android and web app
Experiment complete
Thanks for the feedback!
Description
Trove is retired. Thank you for your usage, we have learned a lot from this experiment.
Visit the Trove website for more on the project retired.
—
Trove, a Microsoft Garage project, is an app for Android and web app that provides people with an opportunity to contribute to AI projects by submitting photos for payment that developers can then use to train machine learning models.
Developers and researchers rely on large batches of high quality and diverse photos to train their computer vision models, and they often crowd source pictures by enlisting the help of gig workers to submit them. Unfortunately, photo contributors can have little to no insight into the greater context or purpose of the project and when photos don’t meet the project’s needs, their work can be rejected without explanation.
Trove solves this need by connecting contributors with AI developers through a new marketplace that puts communication, context, and feedback first.
Key features include
Contributors can get paid for contributing new or existing photos to AI projects (US only)
- View AI projects that need photos and their corresponding photo offer price
- Learn about the purpose of the project and the desired parameters for photo submissions
- Contribute relevant photos to projects you like from your Camera Roll or by taking a new picture
- Get paid for photos accepted by project owners by connecting a PayPal account
- Communicate directly with developers to ask questions or receive feedback
- Review project owners and contribute to a people first community
Developers can request photos by listing AI projects in the app
- Get relevant photos for AI projects
- Generate enthusiasm and higher submission quality for your project by offering payment for approved payments and explaining the greater purpose and benefits of your model
- Provide feedback directly to contributors to help them re-align their submitted photos to meet your project’s parameters
- Review and thank contributors
Trove is rooted in fair trade with no hidden fees and provides a license that will protect the ownership of your images. The result is responsibly sourced data that delivers more choice, control, and security for developers and contributors. Trove builds trusted connections between the people who depend on data and the people who have it, creating a virtual partnership that will ultimately result in shared contribution to higher quality AI experiences. Read more about the core principles of transparency, control, enrichment, and connection that guided the creation of this project in the Original Garage Blog.
Meet the team
Andreas Petsas
Brian Marble
Christian Liensberger
Devis Lucato
Hugh Xiong
Karthik Balasubramanian
Kelsey Baird
Krishnan Raghupathi
Trinh Duong
The Trove Team
Redmond, US