Fetching the paper…

Constrained Output Embeddings for End-to-End Code-Switching Speech Recognition with Only Monolingual Data · Around