检索结果-内蒙古大学图书馆

End-to-end speech separation with unfolded iterative phase reconstruction

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Wang, Zhong-Qiu Le Roux, Jonathan Wang, DeLiang Hershey, John R. United States Department of Computer Science and Engineering Ohio State University United States Center for Cognitive and Brain Sciences Ohio State University United States

This paper proposes an end-to-end approach for single-channel speaker-independent multi-speaker speech separation, where time-frequency (T-F) masking, the short-time Fourier transform (STFT), and its inverse are represented as layers within a deep network. Previous approaches, rather than computing a loss on the reconstructed signal, used a surrogate loss based on the target STFT magnitudes. This ignores reconstruction error introduced by phase inconsistency. In our approach, the loss function is directly defined on the reconstructed signals, which are optimized for best separation. In addition, we train through unfolded iterations of a phase reconstruction algorithm, represented as a series of STFT and inverse STFT layers. While mask values are typically limited to lie between zero and one for approaches using the mixture phase for reconstruction, this limitation is less relevant if the estimated magnitudes are to be used together with phase reconstruction. We thus propose several novel activation functions for the output layer of the T-F masking, to allow mask values beyond one. On the publicly-available wsj0-2mix dataset, our approach achieves state-of-the-art 12.6 dB scale-invariant signal-to-distortion ratio (SI-SDR) and 13.1 dB SDR, revealing new possibilities for deep learning based phase reconstruction and representing a fundamental progress towards solving the notoriously-hard cocktail party problem. Copyright © 2018, The Authors. All rights reserved.

关键词： Iterative methods

Slip Detection with Combined Tactile and Visual Information

学校读者我要写书评

暂无评论

Slip Detection with Combined Tactile and Visual Information

IEEE International Conference on Robotics and Automation (ICRA)

作者： Jianhua Li Siyuan Dong Edward Adelson Computer Science and Artificial Intelligence Laboratory (CSAIL) MIT Cambridge MA USA Department of Electrical Engineering & Computer Science and CSAIL MIT Cambridge MA USA Faculty of Department of Brain and Cognitive Sciences and CSAIL MIT Cambridge MA USA

Slip detection plays a vital role in robotic manipulation and it has long been a challenging problem in the robotic community. In this paper, we propose a new method based on deep neural network (DNN) to detect slip. The training data is acquired by a GelSight tactile sensor and a camera mounted on a gripper when we use a robot arm to grasp and lift 94 daily objects with different grasping forces and grasping positions. The DNN is trained to classify whether a slip occurred or not. To evaluate the performance of the DNN, we test 10 unseen objects in 152 grasps. A detection accuracy as high as 88.03 % is achieved. It is anticipated that the accuracy can be further improved with a larger dataset. This method is beneficial for robots to make stable grasps, which can be widely applied to automatic force control, grasping strategy selection and fine manipulation.

关键词： Grippers Grasping Cameras Tactile sensors Force

brain network construction and classification toolbox (brainnetclass)

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Zhou, Zhen Chen, Xiaobo Zhang, Yu Qiao, Lishan Yu, Renping Pan, Gang Zhang, Han Shen, Dinggang College of Computer Science and Technology Zhejiang University Hangzhou310027 China Department of Radiology and BRIC University of North Carolina at Chapel Hill Chapel HillNC27599 United States Automotive Engineering Research Institute Jiangsu University Zhenjiang212013 China Department of Psychiatry and Behavior Sciences Stanford University StanfordCA94305 United States School of Mathematics Science Liaocheng University Liaocheng252000 China School of Electric Engineering Zhengzhou University Zhengzhou450001 China Department of Brain and Cognitive Engineering Korea University Seoul02841 Korea Republic of

brain functional network has become an increasingly used approach in understanding brain functions and diseases. Many network construction methods have been developed, whereas the majority of the studies still used static pairwise Pearson’s correlation-based functional connectivity. The goal of this work is to introduce a toolbox namely "brain Network Construction and Classification" (brainNetClass) to the field to promote more advanced brain network construction methods. It comprises various brain network construction methods, including some state-of-the-art methods that were recently developed to capture more complex interactions among brain regions along with connectome feature extraction, reduction, parameter optimization towards network-based individualized classification. brainNetClass is a MATLAB-based, open-source, cross-platform toolbox with graphical user-friendly interfaces for cognitive and clinical neuroscientists to perform rigorous computer-aided diagnosis with interpretable result presentations even though they do not possess neuroimage computing and machine learning knowledge. We demonstrate the implementations of this toolbox on real datasets. brainNetClass (v1.0) can be downloaded from https://***/zzstefan/brainNetClass. Copyright © 2019, The Authors. All rights reserved.

关键词： computer aided diagnosis

Towards universal dialogue state tracking

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Ren, Liliang Xie, Kaige Chen, Lu Yu, Kai Key Lab. of Shanghai Education Commission for Intelligent Interaction and Cognitive Eng. SpeechLab Department of Computer Science and Engineering Brain Science and Technology Research Center Shanghai Jiao Tong University Shanghai China

Dialogue state tracking is the core part of a spoken dialogue system. It estimates the beliefs of possible user's goals at every dialogue turn. However, for most current approaches, it's difficult to scale to large dialogue domains. They have one or more of following limitations: (a) Some models don't work in the situation where slot values in ontology changes dynamically;(b) The number of model parameters is proportional to the number of slots;(c) Some models extract features based on hand-crafted lexicons. To tackle these challenges, we propose StateNet, a universal dialogue state tracker. It is independent of the number of values, shares parameters across all slots, and uses pre-trained word vectors instead of explicit semantic dictionaries. Our experiments on two datasets show that our approach not only overcomes the limitations, but also significantly outperforms the performance of state-of-the-art approaches. Copyright © 2018, The Authors. All rights reserved.

关键词： Semantics

Policy Adaptation for Deep Reinforcement Learning-Based Dialogue Management

学校读者我要写书评

暂无评论

Policy Adaptation for Deep Reinforcement Learning-Based Dial...

IEEE International Conference on Acoustics, Speech and Signal Processing

作者： Lu Chen Cheng Chang Zhi Chen Bowen Tan Milica Gasic Kai Yu Key Lab. of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering SpeechLab Department of Computer Science and Engineering Brain Science and Technology Research Center Shanghai Jiao Tong University Shanghai China Engineering Department University of Cambridge Cambridge UK

ISBN: (纸本)9781538646595

Policy optimization is the core part of statistical dialogue management. Deep reinforcement learning has been successfully used for dialogue policy optimization for a static pre-defined domain. However, when the domain changes dynamically, e.g. a new previously unseen concept (or slot) which can be then used as a database search constraint is added, or the policy for one domain is transferred to another domain, the dialogue state space and action sets both will change. Therefore, the model structures for different domains have to be different. This makes dialogue policy adaptation/transfer challenging. Here a multi -agent dialogue policy (MADP) is proposed to tackle these problems. MADP consists of some slot-dependent agents (S-Agents) and a slot-independent agent (G-Agent). S-Agents have shared parameters in addition to private parameters for each one. During policy transfer, the shared parameters in S-Agents and the parameters in G-Agent can be directly transferred to the agents in extended/new domain. Simulation experiments showed that MADP can significantly speed up the policy learning and facilitate policy adaptation.

关键词： dialogue policy deep reinforcement learning adaptation multi-agent Policies Domain Agents learning (artificial intelligence) state space Reinforcement

Soft Realization: a Bio-inspired Implementation Paradigm

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Mahdiani, H.R. Bojnordi, M.N. Fakhraie, S.M. Computer Science and Engineering Department Shahid Beheshti University Tehran Iran Institute for Cognitive and Brain Science Shahid Beheshti University Tehran Iran School of Computing University of Utah Salt Lake City United States Electrical and Computer Engineering Department University of Tehran Tehran Iran

Researchers traditionally solve the computational problems through rigorous and deterministic algorithms called as Hard Computing. Except for some rare analog implementations, these precise algorithms have widely been realized using digital technology as an inherently reliable and accurate implementation platform, either in hardware or software forms. This rigid form of implementation which we refer as Hard Realization relies on strict algorithmic accuracy constraints dictated to digital design engineers. Regardless of the type of application, hard realization admits paying as much as necessary implementation costs to preserve computation precision and determinism throughout all the design and implementation steps. Despite all prior accomplishments in hard realization, this conventional paradigm has encountered serious challenges with today's emerging applications and implementation technologies. Unlike traditional hard computing, the emerging soft and bio-inspired algorithms do not rely on fully precise and deterministic computation, thereby not fully satisfied by old-fashion implementations. Moreover, the incoming nanotechnologies face increasing reliability issues that prevent them from being efficiently exploited in hard realization of applications. This article examines Soft Realization, a novel bio-inspired approach to design and implementation of an important category of applications noticing the internal brain structure. The proposed paradigm mitigates major weaknesses of hard realization by (1) alleviating incompatibilities with today's soft and bio-inspired algorithms such as artificial neural networks, fuzzy systems, and human sense signal processing applications, and (2) resolving the destructive inconsistency with unreliable nanotechnologies. Our experimental results on a set of well-known soft applications implemented using the proposed soft realization paradigm in both reliable and unreliable technologies indicate that significant energy, delay, and a

关键词： Nanotechnology

A lack of efficacy of continuous theta burst stimulation over the left dorsolateral prefrontal cortex in autism: A double blind randomized sham-controlled trial

学校读者我要写书评

暂无评论

Autism research : official journal of the International Society for Autism Research 2023年第6期16卷 2023 May 23页

作者： Hsing-Chang Ni Yi-Lung Chen Yi-Ping Chao Chen-Te Wu Rao-Shayn Chen Tai-Li Chou Susan Shur-Fen Gau Hsiang-Yuan Lin Department of Psychiatry Chang Gung Memorial Hospital at Linkou Taoyuan Taiwan. Department of Healthcare Administration Asia University Taichung Taiwan. Department of Psychology Asia University Taichung Taiwan. Department of Computer Science and Information Engineering Chang Gung University Taoyuan Taiwan. Department of Otorhinolaryngology-Head and Neck Surgery Chang Gung Memorial Hospital at Linkou Taoyuan Taiwan. Department of Medical Imaging and Intervention Chang Gung Memorial Hospital at Linkou Taoyuan Taiwan. Department of Neurology Chang Gung Memorial Hospital at Linkou Taoyuan Taiwan. Department of Psychology National Taiwan University Taipei Taiwan. Graduate Institute of Brain and Mind Sciences College of Medicine National Taiwan University Taipei Taiwan. Neurobiology and Cognitive Science Center National Taiwan University Taipei Taiwan. Department of Psychiatry National Taiwan University Hospital and College of Medicine Taipei Taiwan. Azrieli Adult Neurodevelopmental Centre Campbell Family Mental Health Research Institute Centre for Addiction and Mental Health Toronto Ontario Canada. Department of Psychiatry Temerty Faculty of Medicine University of Toronto Toronto Ontario Canada.

Although previous open-label trials suggest the therapeutic potential of inhibitory repetitive transcranial magnetic stimulation (rTMS) over the dorsolateral prefrontal cortex (DLPFC) in autism spectrum disorder (ASD), methodological caveats exist. We conducted an 8-week randomized, double-blind sham-controlled trial to investigate the efficacy of inhibitory continuous theta burst stimulation (cTBS, a variant of rTMS) over the left DLPFC in individuals with ASD. Sixty children, adolescents and young adults (aged 8-30 years) with ASD without co-occurring intellectual disabilities were randomized to a 16-session 8-week cTBS versus sham stimulation course, with a follow-up 4 weeks after the trial. The Active group was not superior to the Sham group in any clinical or neuropsychological metrics at Week 8 or Week 12. Time effects of 8-week cTBS on symptoms and executive function were remarkable in both Active and Sham groups, with comparable response rates and effect sizes of changes in symptoms/cognition between groups. Our results from a sufficiently powered sample do not endorse the superior efficacy of cTBS over the left DLPFC to the shamed stimulation for children, adolescents and adults with ASD. These findings suggest that earlier positive open-label trial findings may be generalized by generalized/placebo effects. This highlights the urgent need for more rTMS/TBS studies with rigorous trial designs in ASD.

关键词： autism spectrum disorder dorsolateral prefrontal cortex inhibitory paradigm randomized sham-controlled trial repetitive magnetic stimulation theta burst stimulation

Ensemble learning of coarse-grained molecular dynamics force fields with a kernel approach

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Wang, Jiang Chmiela, Stefan Müller, Klaus-Robert Noé, Frank Clementi, Cecilia Rice University Center for Theoretical Biological Physics HoustonTX77005 United States Rice University Department of Chemistry HoustonTX77005 United States Technische Universität Berlin Machine Learning Group Berlin10587 Germany Korea University Department of Brain and Cognitive Engineering Seoul02841 Korea Republic of Max Planck Institute for Informatics Saarbrücken66123 Germany Freie Universität Berlin Department of Mathematics and Computer Science Arnimallee 6 Berlin14195 Germany Freie Universität Berlin Department of Physics Arnimallee 14 Berlin14195 Germany Rice University Department of Physics HoustonTX77005 United States

Gradient-domain machine learning (GDML) is an accurate and efficient approach to learn a molecular potential and associated force field based on the kernel ridge regression algorithm. Here, we demonstrate its application to learn an effective coarse-grained (CG) model from all-atom simulation data in a sample efficient manner. The coarse-grained force field is learned by following the thermodynamic consistency principle, here by minimizing the error between the predicted coarse-grained force and the all-atom mean force in the coarse-grained coordinates. Solving this problem by GDML directly is impossible because coarse-graining requires averaging over many training data points, resulting in impractical memory requirements for storing the kernel matrices. In this work, we propose a data-efficient and memory-saving alternative. Using ensemble learning and stratified sampling, we propose a 2-layer training scheme that enables GDML to learn an effective coarse-grained model. We illustrate our method on a simple biomolecular system, alanine dipeptide, by reconstructing the free energy landscape of a coarse-grained variant of this molecule. Our novel GDML training scheme yields a smaller free energy error than neural networks when the training set is small, and a comparably high accuracy when the training set is sufficiently large. Copyright © 2020, The Authors. All rights reserved.

关键词： Molecular dynamics

A robotic multidimensional directed evolution approach applied to fluorescent voltage reporters (vol 14, pg 352, 2018)

学校读者我要写书评

暂无评论

NATURE CHEMICAL BIOLOGY 2018年第9期14卷 901-901页

作者： Piatkevich, Kiryl D. Jung, Erica E. Straub, Christoph Linghu, Changyang Park, Demian Suk, Ho-Jun Hochbaum, Daniel R. Goodwin, Daniel Pnevmatikakis, Eftychios Pak, Nikita Kawashima, Takashi Yang, Chao-Tsung Rhoades, Jeffrey L. Shemesh, Or Asano, Shoh Yoon, Young-Gyu Freifeld, Limor Saulnier, Jessica L. Riegler, Clemens Engert, Florian Hughes, Thom Drobizhev, Mikhail Szabo, Balint Ahrens, Misha B. Flavell, Steven W. Sabatini, Bernardo L. Boyden, Edward S. Media Lab Massachusetts Institute of Technology (MIT) Cambridge MA USA Howard Hughes Medical Institute Department of Neurobiology Harvard Medical School Boston MA USA Department of Electrical Engineering and Computer Science MIT Cambridge MA USA Harvard-MIT Division of Health Sciences and Technology MIT Cambridge MA USA Simons Center Data Analysis Simons Foundation New York NY USA Department of Mechanical Engineering MIT Cambridge MA USA Janelia Research Campus Howard Hughes Medical Institute Ashburn Virginia USA Picower Institute for Learning & Memory and Department of Brain & Cognitive Sciences MIT Cambridge MA USA Department of Molecular and Cellular Biology and Center for Brain Science Harvard University Cambridge MA USA Department of Neurobiology Faculty of Life Sciences University of Vienna Wien Austria Department of Cell Biology and Neuroscience Montana State University Bozeman Montana USA Department of Biological Physics Eotvos University Budapest Hungary Department of Biological Engineering MIT Cambridge MA USA MIT Center for Neurobiological Engineering MIT Cambridge MA USA Department of Brain and Cognitive Sciences MIT Cambridge MA USA MIT McGovern Institute for Brain Research MIT Cambridge MA USA

We developed a new way to engineer complex proteins toward multidimensional specifications using a simple, yet scalable, directed evolution strategy. By robotically picking mammalian cells that were identified, under a microscope, as expressing proteins that simultaneously exhibit several specific properties, we can screen hundreds of thousands of proteins in a library in just a few hours, evaluating each along multiple performance axes. To demonstrate the power of this approach, we created a genetically encoded fluorescent voltage indicator, simultaneously optimizing its brightness and membrane localization using our microscopy-guided cell-picking strategy. We produced the high-performance opsin-based fluorescent voltage reporter Archon1 and demonstrated its utility by imaging spiking and millivolt-scale subthreshold and synaptic activity in acute mouse brain slices and in larval zebrafish in vivo. We also measured postsynaptic responses downstream of optogenetically controlled neurons in C. elegans.

关键词：