Semantic AI and Knowledge Extraction in Geosciences Without GPT-3

Explore the challenges and solutions of applying Natural Language Processing (NLP) and Machine Learning techniques to the geosciences domain in this 41-minute conference talk from the Toronto Machine Learning Series. Learn why massive language models like GPT-3 and T5 may not perform well in specialized industrial applications, and discover data-centric approaches for knowledge extraction in technical fields. Gain insights into generating domain-specific data for fine-tuning language models in geosciences from Dr. Jérôme Massot, a Machine Learning Geoscientist at Schlumberger with extensive experience in AI and NLP applications in the energy sector.