arXiv subjects

Xi Zhou

Publications and source records attributed to Xi Zhou.

60 records · Page 4Linked to original sources

arXiv2018

Cascaded CNN-resBiLSTM-CTC: An End-to-End Acoustic Model For Speech Recognition

Automatic speech recognition (ASR) tasks are resolved by end-to-end deep learning models, which benefits us by less preparation of raw data, and easier transformation between languages. We propose a novel end-to-end deep learning model architecture namely cascaded CNN-resBiLSTM-CTC. In the proposed model, we add residual blocks in BiLSTM layers to extract sophisticated phoneme and semantic information together, and apply cascaded structure to pay more attention mining information of hard negative samples. By applying both simple Fast Fourier Transform (FFT) technique and n-gram language model (LM) rescoring method, we manage to achieve word error rate (WER) of 3.41% on LibriSpeech test clean corpora. Furthermore, we propose a new batch-varied method to speed up the training process in length-varied tasks, which result in 25% less training time.

eess.AS↗

arXiv2018

Learning Discriminative Features with Multiple Granularities for Person Re-Identification

The combination of global and partial features has been an essential solution to improve discriminative performances in person re-identification (Re-ID) tasks. Previous part-based methods mainly focus on locating regions with specific pre-defined semantics to learn local representations, which increases learning difficulty but not efficient or robust to scenarios with large variances. In this paper, we propose an end-to-end feature learning strategy integrating discriminative information with various granularities. We carefully design the Multiple Granularity Network (MGN), a multi-branch deep network architecture consisting of one branch for global feature representations and two branches for local feature representations. Instead of learning on semantic regions, we uniformly partition the images into several stripes, and vary the number of parts in different local branches to obtain local feature representations with multiple granularities. Comprehensive experiments implemented on the mainstream evaluation datasets including Market-1501, DukeMTMC-reid and CUHK03 indicate that our method has robustly achieved state-of-the-art performances and outperformed any existing approaches by a large margin. For example, on Market-1501 dataset in single query mode, we achieve a state-of-the-art result of Rank-1/mAP=96.6%/94.2% after re-ranking.

cs.CV↗

arXiv2018

Joint 3D Face Reconstruction and Dense Alignment with Position Map Regression Network

We propose a straightforward method that simultaneously reconstructs the 3D facial structure and provides dense alignment. To achieve this, we design a 2D representation called UV position map which records the 3D shape of a complete face in UV space, then train a simple Convolutional Neural Network to regress it from a single 2D image. We also integrate a weight mask into the loss function during training to improve the performance of the network. Our method does not rely on any prior face model, and can reconstruct full facial geometry along with semantic meaning. Meanwhile, our network is very light-weighted and spends only 9.8ms to process an image, which is extremely faster than previous works. Experiments on multiple challenging datasets show that our method surpasses other state-of-the-art methods on both reconstruction and alignment tasks by a large margin.

cs.CV↗

arXiv2015

Superconductivity and Disorder Effect in TlNi$_2$Se$_{2-x}$S$_x$ crystals

After our first discovery of superconductivity (SC) with $T_C$=3.7 K in TlNi$_2$Se$_2$, we grew successfully a series of TlNi$_2$Se$_{2-x}$S$_x$ (0.0 $\leq$ x $\leq$2.0) single crystals. The measurements of resistivity, susceptibility and specific heat were carried out. We found that SC with $T_C$=2.3 K also emerges in TlNi$_2$S$_2$ crystal, which appears to involve heavy electrons with an effective mass $m^*$=13$\sim$25 $m_b$, as inferred from the normal state electronic specific heat and the upper critical field, $H_{C2}(T)$. It was found that the $T_C$ and superconducting volume fraction in TlNi$_2$Se$_{2-x}$S$_x$ crystals changes with the disorder degree induced by the partial substitution of S for Se, which is characterized by the residual resistivity ratio (\textit{RRR}). The effect of the disorder on SC may provide some information for understanding the mechanism of SC in this new Ni-chalcogenide system.

cond-mat.supr-con↗

arXiv2013

Multiband Superconductivity of Heavy Electrons in TlNi2Se2 single crystal

Superconductivity has been first observed in TlNi$_2$Se$_2$ at T$_C$=3.7 K and appears to involve heavy electrons with an effective mass $m^*$=14$\sim$20 $m_b$, as inferred from the normal state electronic specific heat and the upper critical field, H_${C2}$(T). Although the zero-field electronic specific heat data, $C_{es}(T)$, in low temperatures (T < 1/4 T$_C$) can be fitted with a gap BCS model, indicating that TlNi$_2$Se$_2$ is a fully gapped superconductor, the two-gap BCS model presents the best fit to all the $C_{es}(T)$ data below $T_C$. It is also found that the electronic specific heat coefficient in the mixed state, $γ_N(H)$, exhibits a \textit{H}$^{1/2}$ behavior, which was also observed in some \textit{s}-wave superconductors, although once considered as a common feature of the \textit{d}-wave superconductors. Anyway, these results indicate that TlNi$_2$Se$_2$, as a non-magnetic analogue of TlFe$_x$Se$_2$ superconductor, is a multiband superconductor of heavy electron system.

cond-mat.supr-con↗

arXiv2012

MIDI-LAB, a Powerful Visual Basic Program for Creating MIDI Music

Creating MIDI music can be a practical challenge. In the past, working with it was difficult and frustrating to all but the most accomplished and determined. Now, however, we are offering a powerful Visual Basic program called MIDI-LAB, that is easy to learn, and instantly rewarding to even the newest users. MIDI-LAB has been developed to give users the ability to quickly create music with a limitless variety of tunes, tempos, speeds, volumes, instruments, rhythms and major scales. This program has a simple, intuitive, and user-friendly interface, which provides a straightforward way to enter musical data with Numbered Musical Notation (NMN) and immediately create MIDI music. The key feature of this program is the digitalization of music input. It vastly simplifies creating, editing, and saving MIDI music. MIDI-LAB can be used virtually anywhere to write music for entertainment, teaching, computer games, and mobile phone ringtones.

cs.SE↗