Task

Combines retrieval and generation models to find an present information in big corpora. In RAG, a retrieval model first selects relevant pieces of information from a large corpus, which are then used by a generation model to produce a coherent and contextually relevant response. It may or not be Q/A based.

Mathematical representation of text or data that captures the underlying relationships and context within the information. Text embeddings are often used in natural language processing tasks and can also be applied in other areas of machine learning or functional challenges such as search. By transforming data into a dense numerical format, embeddings allow models to better capture the meaning and nuances of the information, leading to more accurate predictions and insights.

Multiple labels can be assigned to a single instance and the goal is to predict all relevant labels for a given instance. This type of problem is commonly used in text or image classification tasks and usually is based on training on a labeled dataset in order to make predictions on new, unlabeled data.