BERT

Transformer-based model developed by google for NLP tasks.

• Core Tasks: text-classification, sentiment-analysis, and Named Entity Recognition (NER).
• Recent Developments:
  • LangExtract: A new open-source Python library from google designed for information extraction from unstructured text using gemini models.
  • Highlights the distinction between traditional NLP methods (like BERT) and the specific challenges of using generative LLMs for non-generative, specific extraction tasks.

Backlink: 2026 04 14 Langextract Sam Witteveen