Foundation Models for Text Generation

Abstract This chapter discusses Foundation Models for Text Generation. includes systems Document Retrieval, which accept a query and return an ordered list of text documents from document collection, often evaluating the similarity embeddings to retrieve relevant passages. Question Answering are given natural language question must provide answer, usually in language. Machine Translation models take one translate it into another Summarization receive long generate short summary covering most important contents document. Generation use autoregressive Language Model longer story, starting initial input. Dialog have task conducting dialog with human partner, typically not limited specific topic.