Monaural source separation is useful for many real-world applications though it is a challenging problem. In this paper, we study deep learning for monaural speech separation. We propose the joint optimization of the deep learning models (deep neural networks and recurrent neural networks) with an extra masking layer, which enforces a reconstruction constraint. Moreover, we explore a discriminative training criterion for the neural networks to further enhance the separation performance. We evaluate our approaches using the TIMIT speech corpus for a monaural speech separation task. Our proposed models achieve about 3.8⇠4.9 dB SIR gain compared to NMF models, while maintaining better SDRs and SARs.

The rapid development of technologies such as tablet PCs and 4G/5G networks has further enhanced the benefits of mobile learning. Although mobile devices are convenient and provide a variety of learning benefits, they are unable to improve students’ learning outcomes without an appropriate learning strategy. Furthermore, little research has been conducted to examine the effects of using collaborative learning on mobile devices. This study proposed a cooperative learning framework using Google Docs to explore the learning outcomes of students of natural science in an elementary curriculum. The study was of a quasi-experimental design with an experimental group (cooperative learning) and a control group (personal learning). The results show that a cooperative learning approach using Google Docs improved learning outcomes, teaching interest, and understanding of campus plants, and reduced cognitive load. One conclusion of the study is that the collaborative learning approach associated with mobile learning is more effective than personal learning. In addition, this paper also provides brief recommendations to expand on the study’s limitations. Future work should investigate the impact of collaborative learning on different environments for mobile learning.

No abstract

No abstract