In this paper, we propose a bilevel joint unsupervised and supervised training (BL-JUST) framework for automatic speech recognition. Compared to the conventional pre-training and fine-tuning strategy which is a disconnected two-stage process, BL-JUST tries to optimize an acoustic model such that it simultaneously minimizes both the unsupervised and supervised loss functions. Because BL-JUST seeks matched local optima of both loss functions, acoustic representations learned by the acoustic model strike a good balance between being generic and task-specific. We solve the BL-JUST problem using penalty-based bilevel gradient descent and evaluate the trained deep neural network acoustic models on various datasets with a variety of architectures and loss functions. We show that BL-JUST can outperform the widely-used pre-training and fine-tuning strategy and some other popular semi-supervised techniques.

BL-JUST employs a lower and upper level optimization In general, bilevel optimization problems are optimization problems with an unsupervised loss and a supervised loss respectively, where the feasible set is determined (in part) using the solution leveraging recent advances in penalty-based bilevel optimization to set of a second optimization problem [10]. Determining the feasible solve this challenging ASR problem with affordable complexity and set is generally called the lower-level problem and the second parametric rigorous convergence guarantees. To evaluate BL-JUST, extensive optimization problem is called the upper-level problem [31, 29].