Weak Supervision
Contents
Weak Supervision#
Weak Supervision is the Data Centric AI technique to
Reduce the efforts of manual labeling while unlocking the vast knowledge of domain subject matter experts (SMEs) by leveraging a diversity of weaker, often programmatic supervision sources.
No Training Data? No Problem! Weak Supervision to the Rescue!#
Presentation at Quantum Black Meetup at Singapore on May 19, 2022: Link to slides
The topics I covered are,
ML’s insatiable need for large datasets
Contemporary ML leaving out domain knowledge from Subject Matter Experts
How Weak Supervision, an approach of Data-Centric AI, solves both the problems simultaneously by encoding domain subject matter expertise into programmatic labeling functions.
The WRENCH benchmark to compare various weak supervision algorithms on several standard datasets.
Snorkel to combine the various labeling functions.
COSINE to fine-tune a final transformer based model that overcomes the noise in weak labels
Future Directions and Resources
The slides have several links to all the necessary resources if you’re interested to read more.
Feel free to use the slides but please remember to credit me with a link back to this repository or my Linkedin profile!
#datascience #machinelearning #artificialintelligence #nlp #algorithms
WeaSEL: Weakly Supervised End-to-end Learning#
An interesting paper/code for those of us working on problems that have low/no labels but high SME domain knowledge. WeaSEL: Weakly Supervised End-to-end Learning.
Train your favorite neural network for weakly-supervised classification:
✅ With only labeling functions (LFs), i.e. without any labeled training data!
✅ In an end-to-end manner, i.e. directly train and evaluate your neural net, there’s no need to train a separate label model anymore as in Snorkel
✅ With a better test set performance and enhanced robustness against correlated or inaccurate LFs than prior methods like Snorkel
🌟 Github: https://github.com/autonlab/weasel
📖 Paper: https://arxiv.org/abs/2107.02233
Have you used any other weak supervision library? Please share in the comments!
#nlp #machinelearning #datascience #researchpaper #github #dataprogramming #weaksupervision