When you attach a conversational text data row to a project, the Labelbox Editor interface will automatically render the conversational text or thread-based text.
For more information on the import format, see our docs on conversational text.
Below are all of the annotation types you may include in your ontology when you are labeling conversational or thread data. Classification-type annotations can be applied at the global level and/or nested within an object-type annotation.
Free-form text classification
To create an entity, simply choose the entity tool in your ontology and select the string of text by clicking the desired starting character and dragging to select a sequence of characters in the unstructured text.
Please note that in both the conversational and thread based UI, the annotator will only be able messages that have been marked as "canLabel" in the import file. The messages that can be annotated will have a white background while the messages that cannot be annotated will have a grey background,
Create an annotation relationship with the following steps:
- Create two entity annotations.
Option+ Click to create a relationship between the entity annotations.
- Define the relationship, selecting from the existing tools in your ontology.
With the annotation relationships feature, you can create and define relationships between entity annotations in unstructured text. You can then use these annotation relationships to consolidate labeling workflows and potentially reduce the number of language models needed.
Below are two common cases for using entity relationships:
Coreference resolution is the task of finding all words/expressions referencing a specific word in a body of text. This is important for NLP tasks that involve natural language understanding, such as document summarization, question answering, and information extraction.
This method usually includes relating nouns, noun phrases, proper nouns, and pronouns. For example, in the phrase, “Federer, the Swiss champion, won again because he served masterfully,” the words Federer, champion, and he are all referring to the same person and should be associated via a relationship.
Dependency parsing is the process of recognizing the grammatical structure of a sentence based on the dependencies between words.
The process establishes relationships between “head” words and their associated dependents. For example, to answer the question “Who is the opponent of Rafael Nadal in his next match?” your model would need to successfully predict the sentence structure to understand that the user wants to know the opponent of Rafael Nadal in the next match.
Create a relationship
Select an entity to be the source of a relationship, then connect it to another entity to be the target.
Updated 14 days ago