แนวคิดหลัก
Extended Flow Matching (EFM) theory extends Flow Matching for conditional generation, ensuring continuity and efficiency in generating distributions.
สถิติ
"We present EFM, an algorithm to learn the matrix field that, through a generalized continuity equation, corresponds to a continuous map from time t and condition c to a probability on the data space."
"EFM performs competitively on the task of conditional generation without guidance-strength-like hyperparameters."
"EFM ensures continuity of the probability distribution with respect to conditions and the generative process."
คำพูด
"EFM aims to match the matrix field as opposed to the vector field."
"Our framework ensures the continuity of the generated conditional distribution through the existence of flow between conditional distributions."