Research on Music Classification Technology Based on Deep Learning

With the advent of the digital music era, digital audio sources have exploded. Music classification (MC) is the basis of managing massive music resources. In this paper, we propose a MC method based on deep learning to improve feature extraction and classifier design based on MIDI (musical instrument digital interface) MC task. Considering that the existing classification technology is limited by the shallow structure, it is difficult for the classifier to learn the time sequence and semantic information of music; this paper proposes a MIDIMC method based on deep learning. In the experiment, we use the MC method proposed in this paper to achieve 90.1% classification accuracy, which is better than the existing classification method based on BP neural network, and verify the music with its classification accuracy. The key point is that the music division method used in this paper has correct MC efficiency. However, due to the limited ability and time involved in the interdisciplinary field, the methodology of this paper has certain limitations, which still needs further research and improvement.

Imaging the Shallow Structure of the Yucca Flat at the Source Physics Experiment Phase II Site with Horizontal-to-Vertical Spectral Ratio Inversion and a Large-N Seismic Array

The Source Physics Experiment (SPE) is a series of chemical explosions at the Nevada National Security Site (NNSS) with the goal of understanding seismic-wave generation and propagation of underground explosions. To understand explosion source physics, accurate geophysical models of the SPE site are needed. Here, we utilize a large-N seismic array deployed at the SPE phase II site to generate a shallow subsurface model of shear-wave velocity. The deployment consists of 500 geophones and covers an area of, approximately, 2.5×2 km. The array is located in the Yucca Flat in the northeast corner of the NNSS, Nye County, Nevada. Using ambient-noise recordings throughout the large-N seismic array, we calculate horizontal-to-vertical spectral ratios (HVSRs) across the array. We obtain 2D seismic images of shear-wave velocities across the SPE phase II site for the shallow structure of the basin. The results clearly image two significant seismic impedance interfaces at ∼150–500 and ∼350–600 m depth. The shallower interface relates to the contrast between Quaternary alluvium and Tertiary volcanic rocks. The deeper interface relates to the contrast between Tertiary volcanic rocks and the Paleozoic bedrock. The 2D subsurface models support and extend previous understanding of the structure of the SPE phase II site. This study shows that the HVSR method in conjunction with a large-N seismic array is a quick and effective method for investigating shallow structures.

Deep multiple non-negative matrix factorization for multi-view clustering

Multi-view clustering aims to group similar samples into the same clusters and dissimilar samples into different clusters by integrating heterogeneous information from multi-view data. Non-negative matrix factorization (NMF) has been widely applied to multi-view clustering owing to its interpretability. However, most NMF-based algorithms only factorize multi-view data based on the shallow structure, neglecting complex hierarchical and heterogeneous information in multi-view data. In this paper, we propose a deep multiple non-negative matrix factorization (DMNMF) framework based on AutoEncoder for multi-view clustering. DMNMF consists of multiple Encoder Components and Decoder Components with deep structures. Each pair of Encoder Component and Decoder Component are used to hierarchically factorize the input data from a view for capturing the hierarchical information, and all Encoder and Decoder Components are integrated into an abstract level to learn a common low-dimensional representation for combining the heterogeneous information across multi-view data. Furthermore, graph regularizers are also introduced to preserve the local geometric information of each view. To optimize the proposed framework, an iterative updating scheme is developed. Besides, the corresponding algorithm called MVC-DMNMF is also proposed and implemented. Extensive experiments on six benchmark datasets have been conducted, and the experimental results demonstrate the superior performance of our proposed MVC-DMNMF for multi-view clustering compared to other baseline algorithms.

Non-native Readers Are More Sensitive to Changes in Surface Linguistic Information than Native Readers*

In the present study we challenge the generally accepted view based primarily on L1 data that surface linguistic information decays rapidly during reading and that only propositional information is retained in memory. In two eye-tracking experiments, we show that both L1 and L2 adult readers retain verbatim information of a text. In particular, the reading behaviour of L2 German learners revealed that they were sensitive to both lexical (synonyms) and syntactic (active/passive alternation) substitutions during a second reading of the texts, while L1 exhibited only reduced sensitivity to the lexical substitutions. The results deliver an important piece of evidence that complies with several current processing (e.g., Shallow Structure Hypothesis), acquisition (Declarative/Procedural Model) and cognitive (e.g., Fuzzy Trace Theory) approaches and adds a new dimension to their empirical and theoretical basis.