Multi-instance learning, as other machine learning tasks, also suffers from the curse of dimensionality. Although dimensionality reduction methods have been investigated for many years, multi-instance dimensionality reduction methods remain untouched. On the other hand, most algorithms in multi-instance framework treat instances in each bag as independently and identically distributed (i.i.d.) samples, which fail to utilize the structure information conveyed by instances in a bag. In this paper, we propose a multi-instance dimensionality reduction method, which treats instances in each bag as non-i.i.d. samples. To capture the structure information conveyed by instances in a bag, we regard every bag as a whole entity. To utilize the bag label information, we maximize the bag margin between positive and negative bags. By maximizing the defined bag margin objective function, we learn a subspace to obtain salient representation of original data. Experiments demonstrate the effectiveness ...