Nathan Susanj

model compression (4) recurrent neural network transducer (3) automatic speech recognition (3) knowledge distillation (2) task-adaptive pretraining (2) adapter module (2) parameter-efficient transfer (2) long short-term memory (2) downstream task (2) speech recognition (1) parameter-efficient transfer learning (1) computational complexity (1) continued pretraining (1) model pruning (1) multi-task learning (1) parameter-efficient learning (1) recurrent neural network (1) spoken language understanding (1) attention mechanism (1) convolutional neural network (1)