Machine learning models—especially large-scale ones like GPT, BERT, or DALL·E—are trained using enormous volumes of data.

Microsoft and partners are building an open toolkit to add natural language interfaces to web content, using the Model ...