Driven by increasing demands of assisting users to dress and match clothing properly, fashion recommendation has attracted wide attention. Its core idea is to model the compatibility among fashion items by jointly projecting embedding into a unified space. However, modeling the item compatibility in such a category-agnostic manner could barely preserve intra-class variance, thus resulting in sub-optimal performance. In this paper, we propose a novel category-aware metric learning framework, which not only learns the cross-category compatibility notions but also preserves the intra-category diversity among items. Specifically, we define a category complementary relation representing a pair of category labels, e.g., tops-bottoms. Given a pair of item embeddings, we first project them to their corresponding relation space, then model the mutual relation of a pair of categories as a relation transition vector to capture compatibility amongst fashion items. We further derive a negative sampling strategy with non-trivial instances to enable the generation of expressive and discriminative item representations. Comprehensive experimental results conducted on two public datasets demonstrate the superiority and feasibility of our proposed approach.
展开▼
