Ambient backscatter (AB) communications and radio frequency (RF)-powered cognitive radio networks (CRNs) address the concerns of energy and spectrum scarcities from different perspectives, and the integration of them has potential benefits for throughput. Motivated by this fact, we study the RF-powered AB-assisted hybrid underlay CRN (ABHU-CRN), and optimize the long-term secondary throughput. Based on the channel states, secondary users choose to perform different actions, and two action spaces are accordingly designed. Due to dynamic and unpredictable environment states, we jointly control the time scheduling and energy management of secondary users, and propose two algorithms, i.e., adjusted-deep deterministic policy gradient (A-DDPG) and combination of A-DDPG and convex optimization (C-ADCO), for the long-term secondary throughput. A-DDPG, a deep reinforcement learning algorithm with continuous spaces, is extended from DDPG to adapt to the design of two action spaces. C-ADCO utilizes the convex optimization that can find the optimal solution to assist A-DDPG to accelerate the convergence. In simulations, the ABHU-CRN under A-DDPG and C-ADCO achieves higher throughput than the optimal throughput of AB-assisted overlay CRN and AB-assisted underlay CRN, which indicates the advantage of the hybrid transmission mode in the ABHU-CRN.

End-to-end Spoken Language Understanding (SLU) models are made increasingly large and complex to achieve the state-ofthe-art accuracy.However, the increased complexity of a model can also introduce high risk of over-fitting, which is a major challenge in SLU tasks due to the limitation of available data.In this paper, we propose an attention-based SLU model together with three encoder enhancement strategies to overcome data sparsity challenge.The first strategy focuses on the transferlearning approach to improve feature extraction capability of the encoder.It is implemented by pre-training the encoder component with a quantity of Automatic Speech Recognition annotated data relying on the standard Transformer architecture and then fine-tuning the SLU model with a small amount of target labelled data.The second strategy adopts multitask learning strategy, the SLU model integrates the speech recognition model by sharing the same underlying encoder, such that improving robustness and generalization ability.The third strategy, learning from Component Fusion (CF) idea, involves a Bidirectional Encoder Representation from Transformer (BERT) model and aims to boost the capability of the decoder with an auxiliary network.It hence reduces the risk of over-fitting and augments the ability of the underlying encoder, indirectly.Experiments on the FluentAI dataset show that cross-language transfer learning and multi-task strategies have been improved by up to 4.52% and 3.89% respectively, compared to the baseline.