The invention provides a man-
machine interaction method and device, equipment and a storage medium, and relates to the technical field of
artificial intelligence, and the method comprises the steps: obtaining multi-
modal data of a current interaction round, the multi-
modal data at least comprising an input image and an input text; based on the input image, performing image recognition by applying a multi-
modal large
language model to obtain image information corresponding to the input image; based on the input text, performing intention recognition by applying a multi-modal large
language model to obtain a retrieval intention corresponding to the input text; based on the image information, the retrieval intention and a historical commodity
retrieval result corresponding to the historical interaction round, performing decision analysis by applying a multi-modal large
language model, and determining a retrieval strategy; the retrieval strategy comprises a joint
knowledge base retrieval strategy, a historical
retrieval result reuse strategy or a non-retrieval strategy; and obtaining reply information based on the retrieval strategy. Differentiated retrieval is performed according to actual retrieval requirements, so that consumption of invalid resources is reduced, and retrieval efficiency is improved.