Introducing New Question Type: Multiple Choice

With the new Multiple Choice question type in Datasaur, you can now assign multiple labels to each row or document, making it easier to capture comprehensive annotations in row-based and document-based labeling projects. Labeling projects often require capturing detailed insights from text, and sometimes a single answer isn’t enough. The Multiple Choice question type resolves this issue perfectly.

Why Multiple Choice?

Many NLP tasks involve scenarios where multiple labels apply to a given text. For example:

• Sentiment Analysis: A sentence might express both positive and negative emotions.
• Intent Recognition: A customer inquiry could involve multiple topics (e.g., billing and technical support).
• Topic Classification: An article might cover multiple themes at once.

Previously, users could use the Dropdown question type with the "allow multiple answers" setting to select multiple labels. However, Multiple Choice provides a more intuitive and structured way to apply multiple labels without relying on dropdown selections, making it a convenient choice for row-based and document-based labeling projects. This improves usability and ensures a smoother annotation experience.

To learn more about multiple-choice or other question types, visit our documentation.