snorkel VS skweak

Hello everyone, we are trying to make a large dataset for Sign Language translation, inspired by BSL-1K [1]. As part of cleaning our collected videos, we use a nice technique for aggregating heuristic labels [2]. We thought it was interesting enough to share with people on here.

[1] https://www.robots.ox.ac.uk/~vgg/research/bsl1k/

[2] https://github.com/snorkel-team/snorkel

The paid product came out of an open source tool: https://github.com/snorkel-team/snorkel

Actually, I referenced it in my issue as well. There seems to be different utils.py file in different folders under the snorkel-tutorials repo but the utils file we get after importing snorkel has a different [file](https://github.com/snorkel-team/snorkel/blob/master/snorkel/utils/core.py) ,i.e. the utils file is different in the main snorkel repo

Snorkel (https://github.com/snorkel-team/snorkel) might provide you exactly what you are looking for. From the docs:

Thanks for pointing me in the right direction. Seems like there’s a few other approaches with weak supervision: https://github.com/NorskRegnesentral/skweak

Code for https://arxiv.org/abs/2104.09683 found: https://github.com/NorskRegnesentral/skweak

Hi Raza here, one of the other co-founders.

I know that HN likes to nerd out over technical details so thought I’d share a bit more on how we aggregate the noisy labels to clean them up.

At the moment we use the great Skweak [1] open source library to do this. Skweak uses an HMM to infer the most likely unobserved label given the evidence of the votes from each of the labelling functions.

This whole strategy of first training a label model and then training a neural net was pioneered by Snorkel. We’ve used this approach for now but we actually think there are big opportunities for improvement.

We’re working on an end-to-end approach that de-noises the labelling function and trains the model at the same time. So far we’ve seen improvements on the standard benchmarks [2] and are planning to submit to Neurips.

R

[1]: Skweak package: https://github.com/NorskRegnesentral/skweak

skweak.

For more advanced unsupervised labeling, you should check skweak

I'm the main developer behind skweak by the way, happy to hear you're interested in our toolkit :-) We do already have a small list of products (see https://github.com/NorskRegnesentral/skweak/blob/main/data/products.json) extracted from DBPedia and Wikidata, but it may not be exactly the type of products you're looking for.