Combining End-to-End and Adversarial Training for Low-Resource Speech Recognition | IEEE Conference Publication | IEEE Xplore