Publicly available software will include tools for the homogenization of highly heterogeneous images (e.g. compound images), for weakly supervised extraction of multimodal semantic concepts from text and images and to visualize, navigate and refine knowledge structures.
