Can we "translate" an English query into a Chinese relevance model? Yes, we just need to have access to a parallel corpus.
