Adaptive Knowledge Distillation Between Text and Speech Pre-Trained Models | IEEE Conference Publication | IEEE Xplore