Filters








9,217 Hits in 3.8 sec

Fuzzy Logic Interpretation of Quadratic Networks [article]

Fenglei Fan, Ge Wang
2019 arXiv   pre-print
how a deep neural network works with a second-order network as the system model.  ...  In each second-order neuron, a quadratic function is used in the place of the inner product in a traditional neuron, and then undergoes a nonlinear activation.  ...  Model Mimicking Methods: Model mimicking methods [12] [13] [14] [15] [16] find the well-performed models that are more interpretable than existing models. For example, Wu et al.  ... 
arXiv:1807.03215v3 fatcat:crfsmcjax5fdxl47asskrmzd5e

Simulation of a class of non-normal random processes

Kurtis R. Gurley, Ahsan Kareem, Michael A. Tognarelli
1996 International Journal of Non-Linear Mechanics  
A neural network system identification model is employed for simulation of output when measured system input is available, and also demonstrates the ability to match higher order spectral characteristics  ...  This method is able to produce simulations which closely match the sample process histogram, power spectral density, and central moments through fourth order.  ...  Acknowledyements-The support for this work was provided in part by ONR grant N00014-93-1-0761, and NSF grants CMS9402196 and CMS95-03779.  ... 
doi:10.1016/0020-7462(96)00025-x fatcat:3zpx2yzlwzbgtomkydvzrsacau

Robust spectrotemporal decomposition by iteratively reweighted least squares

Demba Ba, Behtash Babadi, Patrick L. Purdon, Emery N. Brown
2014 Proceedings of the National Academy of Sciences of the United States of America  
For the neural spiking data, we obtain a new spectral representation of neuronal firing rates.  ...  We assume a Gaussian or a point-process observation model for the time series and introduce prior distributions on the time-frequency plane that yield maximum a posteriori (MAP) spectral estimates that  ...  In principle, these latter models can be applied to time-varying parametric spectral estimation using autoregressive models.  ... 
doi:10.1073/pnas.1320637111 pmid:25468968 pmcid:PMC4273341 fatcat:gnahnfucinevfa2psurmcditeu

Squaring cortex with color

Brian A Wandell, E J Chichilnisky
2012 Nature Neuroscience  
It may be possible to obtain larger and more completely characterized samples using next-generation recording methods.  ...  measurements and Figure 1 Clustering of V1 neuron color tuning estimated using linear and quadratic models of neural response.  ... 
doi:10.1038/nn.3124 pmid:22627792 fatcat:zwhtpmv7dnajnfvtkroph5kfky

Area and Depth Investigation of Anzali Pond Using Satellite Imageries and Group Method of Data Handling Neural Network

Farshad Parhizkar Miandehi
2014 International Journal of Intelligent Information Systems  
This study measures the area of basin surface and predicts the process of changes in the climate of the pond neighborhood during the next years, using GMDH neural network.  ...  The main reason for this reduction is diversion of rivers, sediment entering and changes in land use around the pond.  ...  Acknowledgement The authors would like to thank Guilan weather station for providing meteorological data that have been utilized in this research study.  ... 
doi:10.11648/j.ijiis.s.2014030601.22 fatcat:pudwhpdairdnnl3zzpzcog53vy

Application of Recurrent Neural Network for the Prediction of Target Non-Apneic Arousal Regions in Physiological Signals

Naimahmed Nesaragi, Shubha Majumder, Ashish Sharma, Kouhyar Tavakolian, Shivnarayan Patidar
2018 2018 Computing in Cardiology Conference (CinC)  
This work presents a new method for detection of target non-apneic arousals by applying a recurrent neural network architecture on the various specified polysomnographic (PSG) signals.  ...  As a second stage, some quadratic discriminant (QD) layers are modelled and appended to the trained LSTMs in groups.  ...  signals" and (b) DST India, ECR project entitled "Analysis of cardiovascular disorders using heart sound signals", project no.  ... 
doi:10.22489/cinc.2018.256 dblp:conf/cinc/NesaragiMSTP18 fatcat:25vazc7cc5aedf3b3ek2cwq6mq

Spectral Learning For Expressive Interactive Ensemble Music Performance

Guangyu Xia, Yun Wang, Roger B. Dannenberg, Geoffrey Gordon
2015 Zenodo  
Methods for We use three methods for comparison: regression, neural network, and the timing estimation often used in automatic accompaniment systems [6] .  ...  ., using equal window sizes for history and future), not attempting to give a full description of how to use spectral methods.  ... 
doi:10.5281/zenodo.1415805 fatcat:3n6wid7g6fbela3rk2jaeepa6m

YellowFin and the Art of Momentum Tuning [article]

Jian Zhang, Ioannis Mitliagkas
2018 arXiv   pre-print
YellowFin optionally uses a negative-feedback loop to compensate for the momentum dynamics in asynchronous settings on the fly.  ...  We empirically show that YellowFin can converge in fewer iterations than Adam on ResNets and LSTMs for image recognition, language modeling and constituency parsing, with a speedup of up to 3.28x in synchronous  ...  Acknowledgements We are grateful to Christopher Ré for his valuable guidance and support.  ... 
arXiv:1706.03471v2 fatcat:3auj7p76hra47dnfdigldrrrny

Adaptive stimulus optimization for sensory systems neuroscience

Christopher DiMattina, Kechen Zhang
2013 Frontiers in Neural Circuits  
In this paper, we review several lines of recent work aimed at developing practical methods for adaptive on-line stimulus generation for sensory neurophysiology.  ...  Finally, we present a two-stage stimulus design procedure which combines the dual goals of model estimation and model comparison and may be especially useful for system identification experiments where  ...  MULTIPLE LAYER NEURAL NETWORKS Since many sensory neurons are non-linear Wu et al., 2006) , it is of interest to characterize neurons using various non-linear models, including quadratic and bilinear  ... 
doi:10.3389/fncir.2013.00101 pmid:23761737 pmcid:PMC3674314 fatcat:6g46yiib4zebzkpw2y6fav2rjm

Identifying Complex Brain Networks Using Penalized Regression Methods

Eduardo Martínez-Montes, Mayrim Vega-Hernández, José M. Sánchez-Bornot, Pedro A. Valdés-Sosa
2008 Journal of biological physics (Print)  
In this work we study the application of new penalized regression methods to i) the spatial characterization of the brain networks associated with the identification of faces and ii) the PARAFAC analysis  ...  Multiple penalized least squares model .  ...  Acknowledgments The authors thank Mark Cohen and Jhoanna Pérez-Hidalgo-Gato for kindly providing the data of the resting EEG and face identification experiment used in this study.  ... 
doi:10.1007/s10867-008-9077-0 pmid:19669480 pmcid:PMC2585631 fatcat:ffzk5hjveve6renuxvcqki7omq

Fiedler Regularization: Learning Neural Networks with Graph Sparsity [article]

Edric Tam, David Dunson
2020 arXiv   pre-print
We propose to use the Fiedler value of the neural network's underlying graph as a tool for regularization. We provide theoretical support for this approach via spectral graph theory.  ...  Existing regularization methods often focus on dropping/penalizing weights in a global manner that ignores the connectivity structure of the neural network.  ...  Acknowledgements We would like to thank Julyan Arbel for pointing out a mistake in the initial draft of this paper, now corrected.  ... 
arXiv:2003.00992v3 fatcat:xmiaqs6gufc2bc2pcltcelvype

Robust Estimation of Sparse Narrowband Spectra from Neuronal Spiking Data

Sina Miran, Patrick L. Purdon, Emery N. Brown, Behtash Babadi
2017 IEEE Transactions on Biomedical Engineering  
Although spectral analysis techniques are widely used in the analysis of noninvasive neural recordings such as EEG, their application to spiking data is limited due to the binary and non-linear nature  ...  Application of our method to clinically recorded spiking data from a patient under general anesthesia reveals a striking resemblance between our estimated power spectral density and that of the local field  ...  The authors would like to thank Alireza Sheikhattar for helpful discussions, as well as the anonymous reviewers for their various insightful comments and suggestions.  ... 
doi:10.1109/tbme.2016.2642783 pmid:28026746 pmcid:PMC5665393 fatcat:mxzymdf3b5eppde7osqwzfwaji

Suspended Matter Model On Alsat-1 Image By Mlp Network And Mathematical Morphology: Prototypes By K-Means

S. Loumi, H. Merrad, F. Alilat, B. Sansal
2007 Zenodo  
In this article, we propose a methodology for the characterization of the suspended matter along Algiers-s bay.  ...  The mask which selects the zone of interest (water in our case) was carried out by using a multi spectral classification by ISODATA algorithm.  ...  ACKNOWLEDGMENT The authors wish to thank the ASAL (Algerian Agency Space) Institute for the Alsat-1 images, and the ISMAL institute (Institute of Marine Science of the Algerian Littoral) for in situ measurements  ... 
doi:10.5281/zenodo.1079504 fatcat:kjkqsbwv6zcmhdo2gjjru2pfly

Auditory fMRI of Sound Intensity and Loudness for Unilateral Stimulation [chapter]

Oliver Behler, Stefan Uppenkamp
2016 Advances in Experimental Medicine and Biology  
for various anatomically defined regions of interest in the ascending auditory pathway and in the cortex.  ...  We report a systematic exploration of the interrelation of sound intensity, ear of entry, individual loudness judgments, and brain activity across hemispheres, using auditory functional magnetic resonance  ...  This approach allowed us to characterize the neural representation of sound intensity and loudness in a detailed way.  ... 
doi:10.1007/978-3-319-25474-6_18 pmid:27080657 fatcat:demskuft3rg7lovf4uhh5z27ce

Multidimensional receptive field processing by cat primary auditory cortical neurons

Craig A. Atencio, Tatyana O. Sharpee
2017 Neuroscience  
Standard methods that are often used to characterize multidimensional stimulus selectivity, such as spike-triggered covariance (STC) or maximally informative dimensions (MIDs), are either limited to Gaussian  ...  An information theoretic extension of STC, the maximum noise entropy (MNE) model, can be used with non-Gaussian stimulus distributions to find an arbitrary number of stimulus dimensions.  ...  Thus, this model of neural processing allows us to identify multiple functional inputs for a given neuron.  ... 
doi:10.1016/j.neuroscience.2017.07.003 pmid:28694174 pmcid:PMC5600511 fatcat:qhntw4t4pva2tpzssz7e7ptvqi
« Previous Showing results 1 — 15 out of 9,217 results