Wonders: Multicultural Images and Captions for Inclusive AI
Help us bridge the gap between generic AI in English and multicultural AI for everyone!
We are researchers from TODO universities around the world, and we need your help to collect images and captions of typical aspects from your culture. This data will improve the next generation of AI models for your country and language!
If you enter your your email and a password below, you will be able revise your images and captions (eg. upload the image from your phone and later write the captions from your laptop).
At the end of our data collection campaign, we will use emails and passwords to create a list of all contributors and provide rewards to the best ones. For details, see our website (TODO).
If your country/language is missing, please reach out to us at FIXME@gmail.com.
Please select your country
Wonders: Multicultural Images and Captions for Inclusive AI
We want to collect images and annotations that are relevant for your culture.
- Main concept: Choose the main category and concept for which you want to share data. You can type a new concept if it’s culturally relevant and missing from our list.
- Image: An image where the main concept is clearly visible. The image needs to (1) show the concept in typical contexts where it can be found in your country, and (2) include other objects/people whenever possible. Pictures taken by yourself are preferred. Otherwise, please add a URL for publicly-licensed images from the Web (e.g., from pexels.com or freepik.com). For more details, see “Instructions” below.
- Description: A long description (around 40 words) that provides extensive and detailed visual information. Descriptions must be objective: focus on how you would describe the image to someone who can’t see it, without adding your own opinions/speculations. The text needs to include the main concept and describes the content of the image in detail by including:
- Who? The visual appearance and observable emotions (e.g., “is smiling”) of persons and animals.
- What? The actions performed in the image.
- Where? The setting of the image, including the size, color and relationships between objects.
- (Optional) Other concepts: Tag any other concept that is visible in the image.
See some examples of data we want to collect at the bottom of the page.
For detailed guidelines, click on the “Instructions” button below.
Notes
- By participating in this task, you agree to donate images and captions under CC BY license, which will allow us to share the data with others to improve AI.
- You will be able to update or remove any provided data at any time if you wish so.
- Participants that provide more data, across more concepts and of high quality (see "Instructions" button) will be rewarded (see TODO).
Browse Data
Loading your data, please wait ...