Contrastive language-image pretraining (CLIP)

{
  "id": "MLC-T-OAG-000838",
  "term": "Contrastive language-image pretraining (CLIP)",
  "field": "Oil and gas",
  "definition": "A multimodal machine learning model, known as Contrastive Language-Image Pretraining, trained on a large dataset of images and their corresponding text descriptions from the internet. It learns to connect textual concepts with visual information, making it proficient at tasks like searching for specific equipment in drone footage based on a text query. The model's output is a similarity score between a given image and a text prompt.",
  "abbreviation": "CLIP",
  "subject": "Digital",
  "url": "https://mlchart.com/terminology/oil-and-gas/contrastive-language-image-pretraining-clip/"
}