Searcharxiv⌕ Search

arXiv subjects

Jian Wu

Publications and source records attributed to Jian Wu.

At least 343 records · Page 19Linked to original sources
arXiv2020

Continuous Speech Separation with Conformer

Continuous speech separation plays a vital role in complicated speech related tasks such as conversation transcription. The separation model extracts a single speaker signal from a mixed speech. In this paper, we use transformer and conformer in lieu of recurrent neural networks in the separation system, as we believe capturing global information with the self-attention based method is crucial for the speech separation. Evaluating on the LibriCSS dataset, the conformer separation model achieves state of the art results, with a relative 23.5% word error rate (WER) reduction from bi-directional LSTM (BLSTM) in the utterance-wise evaluation and a 15.4% WER reduction in the continuous evaluation.

eess.AS↗
arXiv2020

DCCRN: Deep Complex Convolution Recurrent Network for Phase-Aware Speech Enhancement

Speech enhancement has benefited from the success of deep learning in terms of intelligibility and perceptual quality. Conventional time-frequency (TF) domain methods focus on predicting TF-masks or speech spectrum, via a naive convolution neural network (CNN) or recurrent neural network (RNN). Some recent studies use complex-valued spectrogram as a training target but train in a real-valued network, predicting the magnitude and phase component or real and imaginary part, respectively. Particularly, convolution recurrent network (CRN) integrates a convolutional encoder-decoder (CED) structure and long short-term memory (LSTM), which has been proven to be helpful for complex targets. In order to train the complex target more effectively, in this paper, we design a new network structure simulating the complex-valued operation, called Deep Complex Convolution Recurrent Network (DCCRN), where both CNN and RNN structures can handle complex-valued operation. The proposed DCCRN models are very competitive over other previous networks, either on objective or subjective metric. With only 3.7M parameters, our DCCRN models submitted to the Interspeech 2020 Deep Noise Suppression (DNS) challenge ranked first for the real-time-track and second for the non-real-time track in terms of Mean Opinion Score (MOS).

eess.AS↗
arXiv2020

An End-to-end Architecture of Online Multi-channel Speech Separation

Multi-speaker speech recognition has been one of the keychallenges in conversation transcription as it breaks the singleactive speaker assumption employed by most state-of-the-artspeech recognition systems. Speech separation is consideredas a remedy to this problem. Previously, we introduced a sys-tem, calledunmixing,fixed-beamformerandextraction(UFE),that was shown to be effective in addressing the speech over-lap problem in conversation transcription. With UFE, an inputmixed signal is processed by fixed beamformers, followed by aneural network post filtering. Although promising results wereobtained, the system contains multiple individually developedmodules, leading potentially sub-optimum performance. In thiswork, we introduce an end-to-end modeling version of UFE. Toenable gradient propagation all the way, an attentional selectionmodule is proposed, where an attentional weight is learnt foreach beamformer and spatial feature sampled over space. Ex-perimental results show that the proposed system achieves com-parable performance in an offline evaluation with the originalseparate processing-based pipeline, while producing remark-able improvements in an online evaluation.

eess.AS↗
arXiv2020

Stellar scattering and the formation of exponential discs in self-gravitating systems

We show, using the N-body code GADGET-2, that stellar scattering by massive clumps can produce exponential discs, and the effectiveness of the process depends on the mass of scattering centres, as well as the stability of the galactic disc. Heavy, dense scattering centres in a less stable disc generate an exponential profile quickly, with a timescale shorter than 1 Gyr. The profile evolution due to scattering can make a near-exponential disc under various initial stellar distributions. This result supports analytic theories that predict the scattering processes always favour the zero entropy gradient solution to the Jeans/Poisson equations, whose profile is a near-exponential. Profile changes are accompanied by disc thickening, and a power-law increase in stellar velocity dispersion in both vertical and radial directions is also observed through the evolution. Close encounters between stars and clumps can produce abrupt changes in stellar orbits and shift stars radially. These events can make trajectories more eccentric, but many leave eccentricities little changed. On average, orbital eccentricities of stars increase moderately with time.

astro-ph.GA↗
arXiv2020

Echoes in Unidirectionally Rotating Molecules

Abstract We report the experimental observation of molecular unidirectional rotation (UDR) echoes, and analyze their origin and behavior both classically and quantum mechanically. The molecules are excited by two time-delayed polarization-twisted ultrashort laser pulses and the echoes are measured by exploding the molecules and reconstructing their spatial orientation from the detected recoil ions momenta. Unlike alignment echoes which are induced by linearly polarized pulses, here the axial symmetry is broken by the twisted polarization, giving rise to molecular unidirectional rotation. We find that the rotation sense of the echo is governed by the twisting sense of the second pulse even when its intensity is much weaker than the intensity of the first pulse. In our theoretical study, we rely on classical phase space analysis and on three-dimensional quantum simulations of the laser-driven molecular dynamics. Both approaches nicely reproduce the experimental results. Echoes in general, and the unique UDR echoes in particular, provide new tools for studies of relaxation processes in dense molecular gases.

physics.optics↗
arXiv2020

Channel-wise Subband Input for Better Voice and Accompaniment Separation on High Resolution Music

This paper presents a new input format, channel-wise subband input (CWS), for convolutional neural networks (CNN) based music source separation (MSS) models in the frequency domain. We aim to address the major issues in CNN-based high-resolution MSS model: high computational cost and weight sharing between distinctly different bands. Specifically, in this paper, we decompose the input mixture spectra into several bands and concatenate them channel-wise as the model input. The proposed approach enables effective weight sharing in each subband and introduces more flexibility between channels. For comparison purposes, we perform voice and accompaniment separation (VAS) on models with different scales, architectures, and CWS settings. Experiments show that the CWS input is beneficial in many aspects. We evaluate our method on musdb18hq test set, focusing on SDR, SIR and SAR metrics. Among all our experiments, CWS enables models to obtain 6.9% performance gain on the average metrics. With even a smaller number of parameters, less training data, and shorter training time, our MDenseNet with 8-bands CWS input still surpasses the original MMDenseNet with a large margin. Moreover, CWS also reduces computational cost and training time to a large extent.

eess.AS↗
arXiv2020

Theory of Subcycle Linear Momentum Transfer in Strong-Field Tunneling Ionization

Interaction of a strong laser pulse with matter transfers not only energy but also linear momentum of the photons. Recent experimental advances have made it possible to detect the small amount of linear momentum delivered to the photoelectrons in strong-field ionization of atoms. We present numerical simulations as well as an analytical description of the subcycle phase (or time) resolved momentum transfer to an atom accessible by an attoclock protocol. We show that the light-field-induced momentum transfer is remarkably sensitive to properties of the ultrashort laser pulse such as its carrier-envelope phase and ellipticity. Moreover, we show that the subcycle resolved linear momentum transfer can provide novel insights into the interplay between nonadiabatic and nondipole effects in strong-field ionization. This work paves the way towards the investigation of the so-far unexplored time-resolved nondipole nonadiabatic tunneling dynamics.

physics.atom-ph↗
arXiv2020

NPU Speaker Verification System for INTERSPEECH 2020 Far-Field Speaker Verification Challenge

This paper describes the NPU system submitted to Interspeech 2020 Far-Field Speaker Verification Challenge (FFSVC). We particularly focus on far-field text-dependent SV from single (task1) and multiple microphone arrays (task3). The major challenges in such scenarios are short utterance and cross-channel and distance mismatch for enrollment and test. With the belief that better speaker embedding can alleviate the effects from short utterance, we introduce a new speaker embedding architecture - ResNet-BAM, which integrates a bottleneck attention module with ResNet as a simple and efficient way to further improve the representation power of ResNet. This contribution brings up to 1% EER reduction. We further address the mismatch problem in three directions. First, domain adversarial training, which aims to learn domain-invariant features, can yield to 0.8% EER reduction. Second, front-end signal processing, including WPE and beamforming, has no obvious contribution, but together with data selection and domain adversarial training, can further contribute to 0.5% EER reduction. Finally, data augmentation, which works with a specifically-designed data selection strategy, can lead to 2% EER reduction. Together with the above contributions, in the middle challenge results, our single submission system (without multi-system fusion) achieves the first and second place on task 1 and task 3, respectively.

eess.AS↗
arXiv2020

Large Scale Subject Category Classification of Scholarly Papers with Deep Attentive Neural Networks

Subject categories of scholarly papers generally refer to the knowledge domain(s) to which the papers belong, examples being computer science or physics. Subject category information can be used for building faceted search for digital library search engines. This can significantly assist users in narrowing down their search space of relevant documents. Unfortunately, many academic papers do not have such information as part of their metadata. Existing methods for solving this task usually focus on unsupervised learning that often relies on citation networks. However, a complete list of papers citing the current paper may not be readily available. In particular, new papers that have few or no citations cannot be classified using such methods. Here, we propose a deep attentive neural network (DANN) that classifies scholarly papers using only their abstracts. The network is trained using 9 million abstracts from Web of Science (WoS). We also use the WoS schema that covers 104 subject categories. The proposed network consists of two bi-directional recurrent neural networks followed by an attention layer. We compare our model against baselines by varying the architecture and text representation. Our best model achieves micro-F1 measure of 0.76 with F1 of individual subject categories ranging from 0.50-0.95. The results showed the importance of retraining word embedding models to maximize the vocabulary overlap and the effectiveness of the attention mechanism. The combination of word vectors with TFIDF outperforms character and sentence level embedding models. We discuss imbalanced samples and overlapping categories and suggest possible strategies for mitigation. We also determine the subject category distribution in CiteSeerX by classifying a random sample of one million academic papers.

cs.CL↗
arXiv2020

Speaker diarization with session-level speaker embedding refinement using graph neural networks

Deep speaker embedding models have been commonly used as a building block for speaker diarization systems; however, the speaker embedding model is usually trained according to a global loss defined on the training data, which could be sub-optimal for distinguishing speakers locally in a specific meeting session. In this work we present the first use of graph neural networks (GNNs) for the speaker diarization problem, utilizing a GNN to refine speaker embeddings locally using the structural information between speech segments inside each session. The speaker embeddings extracted by a pre-trained model are remapped into a new embedding space, in which the different speakers within a single session are better separated. The model is trained for linkage prediction in a supervised manner by minimizing the difference between the affinity matrix constructed by the refined embeddings and the ground-truth adjacency matrix. Spectral clustering is then applied on top of the refined embeddings. We show that the clustering performance of the refined speaker embeddings outperforms the original embeddings significantly on both simulated and real meeting data, and our system achieves the state-of-the-art result on the NIST SRE 2000 CALLHOME database.

eess.AS↗
arXiv2020

Continuous speech separation: dataset and analysis

This paper describes a dataset and protocols for evaluating continuous speech separation algorithms. Most prior studies on speech separation use pre-segmented signals of artificially mixed speech utterances which are mostly \emph{fully} overlapped, and the algorithms are evaluated based on signal-to-distortion ratio or similar performance metrics. However, in natural conversations, a speech signal is continuous, containing both overlapped and overlap-free components. In addition, the signal-based metrics have very weak correlations with automatic speech recognition (ASR) accuracy. We think that not only does this make it hard to assess the practical relevance of the tested algorithms, it also hinders researchers from developing systems that can be readily applied to real scenarios. In this paper, we define continuous speech separation (CSS) as a task of generating a set of non-overlapped speech signals from a \textit{continuous} audio stream that contains multiple utterances that are \emph{partially} overlapped by a varying degree. A new real recorded dataset, called LibriCSS, is derived from LibriSpeech by concatenating the corpus utterances to simulate a conversation and capturing the audio replays with far-field microphones. A Kaldi-based ASR evaluation protocol is also established by using a well-trained multi-conditional acoustic model. By using this dataset, several aspects of a recently proposed speaker-independent CSS algorithm are investigated. The dataset and evaluation scripts are available to facilitate the research in this direction.

cs.SD↗
arXiv2020

UNet 3+: A Full-Scale Connected UNet for Medical Image Segmentation

Recently, a growing interest has been seen in deep learning-based semantic segmentation. UNet, which is one of deep learning networks with an encoder-decoder architecture, is widely used in medical image segmentation. Combining multi-scale features is one of important factors for accurate segmentation. UNet++ was developed as a modified Unet by designing an architecture with nested and dense skip connections. However, it does not explore sufficient information from full scales and there is still a large room for improvement. In this paper, we propose a novel UNet 3+, which takes advantage of full-scale skip connections and deep supervisions. The full-scale skip connections incorporate low-level details with high-level semantics from feature maps in different scales; while the deep supervision learns hierarchical representations from the full-scale aggregated feature maps. The proposed method is especially benefiting for organs that appear at varying scales. In addition to accuracy improvements, the proposed UNet 3+ can reduce the network parameters to improve the computation efficiency. We further propose a hybrid loss function and devise a classification-guided module to enhance the organ boundary and reduce the over-segmentation in a non-organ image, yielding more accurate segmentation results. The effectiveness of the proposed method is demonstrated on two datasets. The code is available at: github.com/ZJUGiveLab/UNet-Version

eess.IV↗
arXiv2020

Privileged Features Distillation at Taobao Recommendations

Features play an important role in the prediction tasks of e-commerce recommendations. To guarantee the consistency of off-line training and on-line serving, we usually utilize the same features that are both available. However, the consistency in turn neglects some discriminative features. For example, when estimating the conversion rate (CVR), i.e., the probability that a user would purchase the item if she clicked it, features like dwell time on the item detailed page are informative. However, CVR prediction should be conducted for on-line ranking before the click happens. Thus we cannot get such post-event features during serving. We define the features that are discriminative but only available during training as the privileged features. Inspired by the distillation techniques which bridge the gap between training and inference, in this work, we propose privileged features distillation (PFD). We train two models, i.e., a student model that is the same as the original one and a teacher model that additionally utilizes the privileged features. Knowledge distilled from the more accurate teacher is transferred to the student to improve its accuracy. During serving, only the student part is extracted and it relies on no privileged features. We conduct experiments on two fundamental prediction tasks at Taobao recommendations, i.e., click-through rate (CTR) at coarse-grained ranking and CVR at fine-grained ranking. By distilling the interacted features that are prohibited during serving for CTR and the post-event features for CVR, we achieve significant improvements over their strong baselines. During the on-line A/B tests, the click metric is improved by +5.0% in the CTR task. And the conversion metric is improved by +2.3% in the CVR task. Besides, by addressing several issues of training PFD, we obtain comparable training speed as the baselines without any distillation.

cs.IR↗
arXiv2020

Audio-visual Recognition of Overlapped speech for the LRS2 dataset

Automatic recognition of overlapped speech remains a highly challenging task to date. Motivated by the bimodal nature of human speech perception, this paper investigates the use of audio-visual technologies for overlapped speech recognition. Three issues associated with the construction of audio-visual speech recognition (AVSR) systems are addressed. First, the basic architecture designs i.e. end-to-end and hybrid of AVSR systems are investigated. Second, purposefully designed modality fusion gates are used to robustly integrate the audio and visual features. Third, in contrast to a traditional pipelined architecture containing explicit speech separation and recognition components, a streamlined and integrated AVSR system optimized consistently using the lattice-free MMI (LF-MMI) discriminative criterion is also proposed. The proposed LF-MMI time-delay neural network (TDNN) system establishes the state-of-the-art for the LRS2 dataset. Experiments on overlapped speech simulated from the LRS2 dataset suggest the proposed AVSR system outperformed the audio only baseline LF-MMI DNN system by up to 29.98\% absolute in word error rate (WER) reduction, and produced recognition performance comparable to a more complex pipelined system. Consistent performance improvements of 4.89\% absolute in WER reduction over the baseline AVSR system using feature fusion are also obtained.

eess.AS↗
arXiv2019

Query Auto Completion for Math Formula Search

Query Auto Completion (QAC) is among the most appealing features of a web search engine. It helps users formulate queries quickly with less effort. Although there has been much effort in this area for text, to the best of our knowledge there is few work on mathematical formula auto completion. In this paper, we implement 5 existing QAC methods on mathematical formula and evaluate them on the NTCIR-12 MathIR task dataset. We report the efficiency of retrieved results using Mean Reciprocal Rank (MRR) and Mean Average Precision(MAP). Our study indicates that the Finite State Transducer outperforms other QAC models with a MRR score of $0.642$.

cs.IR↗
arXiv2019

Measurement of Beam-Correlated Background Neutrons from the Fermilab Booster Neutrino Beam in ANNIE Phase-I

The Accelerator Neutrino Neutron Interaction Experiment (ANNIE) aims to make a unique measurement of neutron yield from neutrino-nucleus interactions and to perform R&D for the next generation of water-based neutrino detectors. In this paper, we characterize beam-induced neutron backgrounds in the experimental hall at Fermi National Accelerator Laboratory. It is shown that the background levels are sufficiently low to allow the next stage of the experiment to proceed. These measurements are relevant to other Booster Neutrino Beam (BNB) experiments located adjacent to ANNIE Hall, where dirt neutrons and sky-shine could present similar backgrounds.

physics.ins-det↗
arXiv2019

Method and Dataset Mining in Scientific Papers

Literature analysis facilitates researchers better understanding the development of science and technology. The conventional literature analysis focuses on the topics, authors, abstracts, keywords, references, etc., and rarely pays attention to the content of papers. In the field of machine learning, the involved methods (M) and datasets (D) are key information in papers. The extraction and mining of M and D are useful for discipline analysis and algorithm recommendation. In this paper, we propose a novel entity recognition model, called MDER, and constructe datasets from the papers of the PAKDD conferences (2009-2019). Some preliminary experiments are conducted to assess the extraction performance and the mining results are visualized.

cs.LG↗