A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is
XIA WANG received the Ph.D. degree in automation from the China University of Mining and Technology, in 1999. ... HAILIN WANG received the B.S. degree in optoelectronics information science and engineering from the Beijing Institute of Technology, Beijing, China, in 2018. ...doi:10.1109/access.2020.2971423 fatcat:3nxxaqy4bfc23jtsaow77zh4xi
While recent progresses in neural network approaches to single-channel speech separation, or more generally the cocktail party problem, achieved significant improvement, their performance for complex mixtures is still not satisfactory. In this work, we propose a novel multi-channel framework for multi-talker separation. In the proposed model, an input multi-channel mixture signal is firstly converted to a set of beamformed signals using fixed beam patterns. For this beamforming, we propose toarXiv:1803.10924v1 fatcat:tfijy4ujn5cjxjcuga73mhcvci
more »... e differential beamformers as they are more suitable for speech separation. Then each beamformed signal is fed into a single-channel anchored deep attractor network to generate separated signals. And the final separation is acquired by post selecting the separating output for each beams. To evaluate the proposed system, we create a challenging dataset comprising mixtures of 2, 3 or 4 speakers. Our results show that the proposed system largely improves the state of the art in speech separation, achieving 11.5 dB, 11.76 dB and 11.02 dB average signal-to-distortion ratio improvement for 4, 3 and 2 overlapped speaker mixtures, which is comparable to the performance of a minimum variance distortionless response beamformer that uses oracle location, source, and noise information. We also run speech recognition with a clean trained acoustic model on the separated speech, achieving relative word error rate (WER) reduction of 45.76\%, 59.40\% and 62.80\% on fully overlapped speech of 4, 3 and 2 speakers, respectively. With a far talk acoustic model, the WER is further reduced.
The training phases of Deep neural network (DNN) consumes enormous processing time and energy. Compression techniques utilizing the sparsity of DNNs can effectively accelerate the inference phase of DNNs. However, it can be hardly used in the training phase because the training phase involves dense matrix-multiplication using General Purpose Computation on Graphics Processors (GPGPU), which endorse regular and structural data layout. In this paper, we propose the Approximate Random Dropout thatarXiv:1805.08939v2 fatcat:dyhsbg2xkvdvvmcwkz2ynrch54
more »... replaces the conventional random dropout of neurons and synapses with a regular and predefined patterns to eliminate the unnecessary computation and data access. To compensate the potential performance loss we develop a SGD-based Search Algorithm to produce the distribution of dropout patterns. We prove our approach is statistically equivalent to the previous dropout method. Experiments results on MLP and LSTM using well-known benchmarks show that the proposed Approximate Random Dropout can reduce the training time by 20%-77% (19%-60%) when dropout rate is 0.3-0.7 on MLP (LSTM) with marginal accuracy drop.
Lecture Notes in Computer Science
A staged computation is a computation organized in a cascade of stages: each stage produces code for its successive stage; the final stage produces the desired output. An off-line procedure called binding time analysis (BTA) is often used to pre-convert unstaged code into staged code, i.e., code annotated with stage labels, which can guide online staged computation. For dynamic re-optimization purposes, it is advantageous for the order of stages in the cascade to change during runtime; however,doi:10.1007/978-3-540-39815-8_5 fatcat:aesqwwszkjghrppaftepihjwkm
more »... the staged code may not support all permutations of stage sequences. Thus, it is both a and practical question to efficiently decide whether a specific stage sequence is valid for a staged code. Our approach is to encode the set of valid stage sequences for a staged code off-line in a stage ordering language (SOL) to facilitate fast online decision. Contrary to the intuition that we only need a single generic SOL (such as the language of posets of stage labels) to sufficiently and efficiently encode the set of valid stage sequences for any staged code in any staged language, we may need different SOLs for different staged languages. We analyze several staged languages and then present a metatheory on validating a SOL for a given staged language. Our result reveals the relationship between SOLs and semantic properties of staged languages, and can influence the design of staged languages and BTA.
In recent years, machine learning researchers have focused on methods to construct flexible and interpretable prediction models. However, an interpretability evaluation, a relationship between generalization performance and an interpretability of the model and a method for improving the interpretability have to be considered. In this paper, a quantitative index of the interpretability is proposed and its rationality is proved, and equilibrium problem between the interpretability and thearXiv:1811.10469v2 fatcat:twlqizl4mfgkrdm75stwxbmqda
more »... zation performance is analyzed. Probability upper bound of the sum of the two performances is analyzed. For traditional supervised kernel machine learning problem, a universal learning framework is put forward to solve the equilibrium problem between the two performances. The condition for global optimal solution based on the framework is deduced. The learning framework is applied to the least-squares support vector machine and is evaluated by some experiments.
Lecture Notes in Computer Science
Trees are ubiquitous in natural environment and realistic models of tree are also indispensable in computer graphics and virtual reality domains. However, their complexity in geometry and topology make it a great challenge for photo-realistic tree reconstruction. Since tree trunk is the preliminary structure of trees, its modeling is a critical step which plays an important role in tree modeling. Many existing methods focus on the overall resemblance of tree branches but omit the local geometrydoi:10.1007/978-3-642-31439-1_23 fatcat:qyvik6mxizdgre7zzhw5nl45jm
more »... details. In this paper, we perform unilateral scanning of real-world trees and propose an approach that could reconstruct trees from incomplete point clouds. The core of our method contains four parts: local optimal segmentation of tree branch, skeletal point and lines extraction from unilateral branch, the cross-section construction of tree branch, and final tree branch surface generation. Experimental results demonstrate the effectiveness and robustness of our method which could keep realistic shape of trees.
Code search is a common practice for developers during software implementation. The challenges of accurate code search mainly lie in the knowledge gap between source code and natural language (i.e., queries). Due to the limited code-query pairs and large code-description pairs available, the prior studies based on deep learning techniques focus on learning the semantic matching relation between source code and corresponding description texts for the task, and hypothesize that the semantic gaparXiv:2105.09630v1 fatcat:jvbhpyr4ajax5k73p4jpjh2l5y
more »... tween descriptions and user queries is marginal. In this work, we found that the code search models trained on code-description pairs may not perform well on user queries, which indicates the semantic distance between queries and code descriptions. To mitigate the semantic distance for more effective code search, we propose QueCos, a Query-enriched Code search model. QueCos learns to generate semantic enriched queries to capture the key semantics of given queries with reinforcement learning (RL). With RL, the code search performance is considered as a reward for producing accurate semantic enriched queries. The enriched queries are finally employed for code search. Experiments on the benchmark datasets show that QueCos can significantly outperform the state-of-the-art code search models.
Time-Spatial data plays a crucial role for different fields such as traffic management. These data can be collected via devices such as surveillance sensors or tracking systems. However, how to efficiently an- alyze and visualize these data to capture essential embedded pattern information is becoming a big challenge today. Classic visualization ap- proaches focus on revealing 2D and 3D spatial information and modeling statistical test. Those methods would easily fail when data become mas-arXiv:1707.02554v1 fatcat:6lgbbvoeyvbybp3ptdfy3yne34
more »... Recent attempts concern on how to simply cluster data and perform prediction with time-oriented information. However, those approaches could still be further enhanced as they also have limitations for han- dling massive clusters and labels. In this paper, we propose a visualiza- tion methodology for mobility data using artificial neural net techniques. This method aggregates three main parts that are Back-end Data Model, Neural Net Algorithm including clustering method Self-Organizing Map (SOM) and prediction approach Recurrent Neural Net (RNN) for ex- tracting the features and lastly a solid front-end that displays the results to users with an interactive system. SOM is able to cluster the visiting patterns and detect the abnormal pattern. RNN can perform the predic- tion for time series analysis using its dynamic architecture. Furthermore, an interactive system will enable user to interpret the result with graph- ics, animation and 3D model for a close-loop feedback. This method can be particularly applied in two tasks that Commercial-based Promotion and abnormal traffic patterns detection.
PbS is a latent substitute of PbTe thermoelectric materials, which is on account of its superiority in low cost and earth abundance. Here, the thermoelectric transport properties of p-type PbS by doping alkali metals (Na and Li) are investigated and it is verified that Li is a more effective dopant than Na. By introducing Li, the electrical and thermal transport properties were optimized collectively. The electrical transport properties were boosted remarkably via adjusting carrierdoi:10.34133/2020/4084532 pmid:33623904 pmcid:PMC7877382 fatcat:il2utlgrcrhsbhesvvfksg3iri
more »... , and the maximum power factor (PFmax) of ~11.5 μW/cmK2 and average power factor (PFave) ~9.9 μW/cmK2 between 423 and 730 K in Pb0.99Li0.01S were achieved, which are much higher than those (~9.5 and ~7.7 μW/cmK2) of Pb0.99Na0.01S. Doping Li and Na can weaken the lattice thermal conductivity effectively. Combining the enlarged PF with suppressed total thermal conductivity, a maximum ZT ~0.5 at 730 K and a large average ZT ~0.4 at 423-730 K were obtained in p-type Pb0.99Li0.01S, which are higher than ~0.4 and ~0.3 in p-type Pb0.99Na0.01S, respectively.
Wenxuan Wang, Zhenghao Yang, and Yue Wang performed data analysis and draft the manuscript. Hongyu Gao, Yan Wang, and Qiong Zhang provided critical revisions to the article. ... Authors' Contributions Wenxuan Wang, Zhenghao Yang, Yue Wang, Hongyu Gao, Yan Wang, and Qiong Zhang have directly participated in the planning, execution, and analysis of this review. ...doi:10.1155/2021/3278351 pmid:34532500 pmcid:PMC8438587 fatcat:qpmu5p3aqfesxowja4mmquo5t4
., 2007) , cerebellum (Wang et al., 2008) , or white matter (Koubeissi et al., 2013) , reduced seizure severity in TLE. ... Dysfunctions of the EC are frequently observed in epileptic brains, including atrophy (Bartolomei et al., 2005) , hypometabolism (Goffin et al., 2009; Guo et al., 2009; Wang et al., 2014) and cell loss ...doi:10.1016/j.ebiom.2016.11.027 pmid:27908611 pmcid:PMC5161446 fatcat:q2e5sooaw5akrk6sqcnbqjyyci
Author Contributions Conceptualization: Zhenghao Wang, Yunjin Bai. ... Z, Bai Y, Wang J (2020) Effects of diuretic administration on outcomes of extracorporeal shockwave lithotripsy: A systematic review and meta-analysis. ... invasive procedure with PLOS ONE PLOS ONE | https://doi.org/10.1371/journal.pone. 0230059 March 5, 2020 1 / 8 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Wang ...doi:10.1371/journal.pone.0230059 pmid:32134993 fatcat:qa66lkjhnzaotfb2z7i77czsu4
Wang et al. designed a gold(I)-isocyanide complex demonstrating interesting tricolor mechanochromism feature (Wang et al., 2018) . ... ., 2018; Wang et al., 2021; Jackson et al., 2021; Yin et al., 2021; Min et al., 2022) . ...doi:10.3389/fchem.2022.865198 pmid:35308787 pmcid:PMC8931262 fatcat:bhh53cmgmbh75kzxibfowmvpju
: Writing -Original Draft preparation; Fubo Cao: Investigation; Zhenghao Zou: Writing -Review & Editing; Chunguang Wang: Visualization; Jun Wang: Supervision; Haihe Yi: Methodology, Formal analysis. ... Wang et al. ...doi:10.1515/rams-2021-0018 fatcat:odw3gpn7obf5peaqc2zq4u6fyq
Non-local neural networks were proposed by Wang et al. in one of the important works on attention mechanisms in the field of computer vision  . ...doi:10.3390/s21093031 pmid:33925864 fatcat:weapqf6fu5b7dl2abvksam3v24
« Previous Showing results 1 — 15 out of 303 results