Create text classifiers in a few hours using the open-source Label Sleuth system
summary
Claims a usable Label Sleuth no‑code system that combines an intuitive UI, active learning and automated background model training so domain experts build text classifiers in about six hours and lawyers cut review time by 80% time saved. The open‑source tool actually trains models (SVM, Naive Bayes, transformer options), surfaces suspicious labels and contradicting pairs, exports models or raw labels, handles skew, and supports customization and multilingual additions — demonstrated working on long contracts and Wikipedia sentences.
AI-written summary · about
chapters (6)
- 0:00Label Sleuth intro
- 6:46Live demo
- 16:57Efficient model building
- 25:07Advanced customization
- 27:10Summary and resources
- 28:15Q&A
related talks
22mImplementing a Text Classifier based on Lucene & LIBSVM
04mLabel Studio
31mWhen humans teach machines
30mEfficient data labelling with weak supervision
59mWorkshop: Choosing the Best Open Source LLM for Your Application
37mCreating Knowledgebases from unstructured text
30mData Driven Security
20mAn Innovative Approach to Labeling Ground Truth in Speech