Skip to content

LabelModel with LFs dependencies #1596

Open
@chaturv3di

Description

This pertains to Issue #1462 and #1513.

As of v0.9.5, LabelModel training assumes that LFs are conditionally independent given Y.

Are there any plans to update the loss function to reflect the algorithm described in multi-task weak supervision paper?

image

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions