Open naokiabe opened 3 years ago
:wave: Hi! This issue has been marked stale due to inactivity. If no further activity occurs, it will automatically be closed in 14 days.
not stale or done.
:wave: Hi! This issue has been marked stale due to inactivity. If no further activity occurs, it will automatically be closed in 14 days.
Background
Solution
DS-MVP-V3 An extended machine learning model that can learn to detect racially biased expressions in context based on input lableled data, without having to specify the "expression" and "context" separately in the input text. Here, the labeled data consists of <text, classification> pairs and the trained model is to output <expression(s), classification> on a new test text, where expression(s) are sub-expression(s) of the input text that are identified to be biased expressions in context of the rest of the text.
Tasks
Acceptance Criteria
Technical Requirements
Roles Required