Spectrum is now read-only. Learn more about the decision in our official announcement.

snorkel

Building and managing training datasets for machine learning

Channels

# announcements

Team

Jason Alan Fries

Henry Ehrenberg

@manas-joglekar

@braden-hancock

Posts

Members

Info

Using Snorkel for N-ary relations extraction

Is it possible to use Snorkel for N-ary relations extraction?

0

Does LabelModel accept large number of classes?

The LabelModel seems can't generate anything with large number of classes( n=12). is there any constraint with the number of classes in this algorithm? can't figure out why.

2

Snorkel v/s Snorkel DryBell v/s Snorkel MeTaL

I am slightly confused between the 3 variants of Snorkel available out there, can anyone briefly point out the difference between the 3, especially Snorkel v/s Snorkel DryBell?

1

Conditional Probability

I tried the API get_conditional_probs() that returns the list of CP for each LF. If my understanding is right, result[0][0] returns the conditional probability when K =-1 (abstain) and Y=0 (in binary category in my case) . Could you explain how to use the weights to evaluate LF?…

0

How to determine the weights of LFs?

When I applied model's get_weights() API the results show some LFs are weighted as 0 and one LF is weighted as high as .93 . and all weights can't added into 1. I'm not sure how to interpret the API? Could you explain how Snorkel implemented this at high level?

0

Estimation of label model with probabilistic labelling functions

I'm currently using Snorkel for a sequence labelling task, and I wonder whether Snorkel could be adapted to use labelling functions that provide probabilistic distributions over labels instead of single labels (or abstain). Looking at the label model in label_model.py, it seems…

1