6,928 Hits in 9.3 sec

Learning kernels from biological networks by maximizing entropy

K. Tsuda, W. S. Noble
2004 Bioinformatics  
This technique has been used successfully, in conjunction with kernel-based learning methods, to draw inferences from several types of biological networks.  ...  Results: We show that computing the diffusion kernel is equivalent to maximizing the von Neumann entropy, subject to a global constraint on the sum of the Euclidean distances between nodes.  ...  A kernel matrix is identified by maximizing the von Neumann entropy subjected to a set of constraints derived from a network.  ... 
doi:10.1093/bioinformatics/bth906 pmid:15262816 fatcat:nqi5l2jmx5gd7g2a3l3wto4upy

Kernel-based topographic map formation achieved with an information-theoretic approach

Marc M. Van Hulle
2002 Neural Networks  
The result is joint entropy maximization of the kernel outputs, which we adopt as our learning criterion.  ...  A new information-theoretic learning algorithm is introduced for kernel-based topographic map formation.  ...  Our joint entropy maximization algorithm differs from the STVQ algorithm in at least three ways.  ... 
doi:10.1016/s0893-6080(02)00077-1 pmid:12416692 fatcat:ty2bqw6qerb33kiflrgwop5rm4

Neural Network Classification Using Error Entropy Minimization [chapter]

Jorge M. Santos, Luís A. Alexandre, Joaquim Marques de Sá
2005 Biological and Artificial Intelligence Environments  
One way of using the entropy criteria in learning systems is to minimize the entropy of the error between two variables: typically, one is the output of the learning system and the other is the target.  ...  In regression, this problem is solved by making a shift of the final result such that it's average equals the average value of the desired target.  ...  Introduction Since the introduction by Shannon [8] of the concept of entropy, and the posterior generalization made by Renyi [7] , that entropy and information theory concepts have been applied in learning  ... 
doi:10.1007/1-4020-3432-6_34 dblp:conf/wirn/SantosAS04 fatcat:ptq5uptin5ab5ml34ew4inlpvy

Leveraging domain information to restructure biological prediction

Xiaofei Nan, Gang Fu, Zhengdong Zhao, Sheng Liu, Ronak Y Patel, Haining Liu, Pankaj R Daga, Robert J Doerksen, Xin Dang, Yixin Chen, Dawn Wilkins
2011 BMC Bioinformatics  
The goal of this research is to develop an algorithm to identify discrete or categorical attributes that maximally simplify the learning task.  ...  To avoid high computational cost, we approximate the solution by the expected minimum conditional entropy with respect to random projections.  ...  The cannabinoid receptor subtypes CB1 and CB2 data source was from the project supported in part by USA National Institutes of Health (NIH) Grant Number 5P20RR021929 from the National Center for Research  ... 
doi:10.1186/1471-2105-12-s10-s22 pmid:22166097 pmcid:PMC3236845 fatcat:n5oi3eovwrag3k4g5km5c4ctde

A Deep 2-Dimensional Dynamical Spiking Neuronal Network for Temporal Encoding trained with STDP [article]

Matthew Evanusa and Cornelia Fermuller and Yiannis Aloimonos
2020 arXiv   pre-print
Spiking Neural Networks (SNNs) can be trained using biologically-realistic learning mechanisms, and can have neuronal activation rules that are biologically relevant.  ...  We analyze the network in terms of network entropy as a metric of information transfer.  ...  Acknowledgments This work was supported partly by the University of Maryland COMBINE program and NSF award DGE-1632976. The authors would like to also thank Mrs.  ... 
arXiv:2009.00581v1 fatcat:elshjencaffilgratq2uh5ran4

Entropic Spectral Learning for Large-Scale Graphs [article]

Diego Granziol, Binxin Ru, Stefan Zohren, Xiaowen Dong, Michael Osborne, Stephen Roberts
2019 arXiv   pre-print
, and real-world networks, such as the social networks for Orkut, YouTube, and Amazon from the SNAP dataset.  ...  We demonstrate its effectiveness and superiority over existing approaches in learning graph spectra, via experiments on both synthetic networks, such as the Erdős-Rényi and Barabási-Albert random graphs  ...  et al., 2007b , biological networks (Palla et al., 2005) and technological networks.  ... 
arXiv:1804.06802v2 fatcat:2iax5gllfrfuplsatbl6kn77ae

Information theory related learning

Thomas Villmann, José C. Príncipe, Andrzej Cichocki
2011 The European Symposium on Artificial Neural Networks  
It reviews and highlights recent developments and new direction in information related learning, which is a fastly developing research area.  ...  These algorithms are based on the fundamental principles of information theory and relate them implicitly or explicitly to learning algoithms and strategies.  ...  This can be realized by explicit maximization of the respective mutual information [57] , or by learning of appropriate feature transformation optimizing the mutual information based on Rényi-entropies  ... 
dblp:conf/esann/VillmannPC11 fatcat:zg4rg6h3ajcy5klcwqenruzpzi

Deep Deterministic Independent Component Analysis for Hyperspectral Unmixing [article]

Hongming Li, Shujian Yu, Jose C. Principe
2022 arXiv   pre-print
Using the matrix-based Rényi's α-order entropy functional, our network can be directly optimized by stochastic gradient descent (SGD), without any variational approximation or adversarial training.  ...  We develop a new neural network based independent component analysis (ICA) method by directly minimizing the dependence amongst all extracted components.  ...  It maximizes joint entropy of predicted components as an alternative.  ... 
arXiv:2202.02951v2 fatcat:6j7nnhinwzhwpkarsvib6jbxvu

Synergies between Intrinsic and Synaptic Plasticity Based on Information Theoretic Learning

Yuke Li, Chunguang Li, Eleni Vasilaki
2013 PLoS ONE  
From an information-theoretical perspective, the error-entropy minimization (MEE) algorithm has newly been proposed as an efficient training method.  ...  IP is sometimes thought to be an information-maximization mechanism. However, it is still unclear how IP affects the performance of artificial neural networks in supervised learning applications.  ...  This batch version of the information-maximization rule can be derived directly from the objective of entropy maximization (''online'' equivalent).  ... 
doi:10.1371/journal.pone.0062894 pmid:23671642 pmcid:PMC3650036 fatcat:bujszdvk7vfz5fvnm3dznwrcui

Kernel-blending connection approximated by a neural network for image classification

Xinxin Liu, Yunfeng Zhang, Fangxun Bao, Kai Shao, Ziyi Sun, Caiming Zhang
2020 Computational Visual Media  
A kernel mapping connection structure, guaranteed by the function approximation theorem, is devised to blend feature extraction and feature classification through neural network learning.  ...  This paper proposes a kernel-blending connection approximated by a neural network (KBNN) for image classification.  ...  Acknowledgements This work was supported in part by the National Natural Science Foundation of China (Grant Nos. 61972227 and 61672018), the Natural Science  ... 
doi:10.1007/s41095-020-0181-9 fatcat:x3tnrfajhfhxrkuumstryy2ade

Kernel Analysis for Estimating the Connectivity of a Network with Event Sequences

Taro Tezuka, Christophe Claramunt
2017 Journal of Artificial Intelligence and Soft Computing Research  
Estimating the connectivity of a network from events observed at each node has many applications.  ...  The proposed method was evaluated using synthetic and real data, by comparing with methods using transfer entropy and the Victor-Purpura distance.  ...  Acknowledgment This work was supported in part by JSPS KAK-ENHI Grant Numbers 21700121, 25280110, and 25540159.  ... 
doi:10.1515/jaiscr-2017-0002 fatcat:wvrnj4hb6ng3vf322qjlflzmua

Gaussian Filter in CRF Based Semantic Segmentation [article]

Yichi Gu, Qisheng Wu, Jing Li, Kai Cheng
2017 arXiv   pre-print
In this paper, we introduce a multi-resolution neural network for FCN and apply Gaussian filter to the extended CRF kernel neighborhood and the label image to reduce the oscillating effect of CRF neural  ...  Artificial intelligence is making great changes in academy and industry with the fast development of deep learning, which is a branch of machine learning and statistical learning.  ...  Deep Learning Neural Network Artificial Neural Network ANN is a mathematical model of network approximation system inspired by the biological neural networks.  ... 
arXiv:1709.00516v1 fatcat:yjfivat6yjbelknyx5neb3e56a

Towards experimental design using a Bayesian framework for parameter identification in dynamic intracellular network models

Andrei Kramer, Nicole Radde
2010 Procedia Computer Science  
We propose to choose the optimal experiments with respect to identifiability of model parameters by maximizing the information content of the expected outcome, measured as the entropy of the posterior  ...  We introduce a framework for the experimental design problem to infer parameters from steady state observations of intracellular networks.  ...  The parameter θ 9 describes the maximal DAG production rate from conversion of ceramide, which is triggered by a sphingomyelin synthase.  ... 
doi:10.1016/j.procs.2010.04.184 fatcat:2l33kqnf4bbcxi2hd6aymohgcy

A Maximum Entropy approach to Massive Graph Spectra [article]

Diego Granziol, Robin Ru, Stefan Zohren, Xiaowen Dong, Michael Osborne, Stephen Roberts
2019 arXiv   pre-print
Graph spectral techniques for measuring graph similarity, or for learning the cluster number, require kernel smoothing.  ...  Our method's computational cost is linear in the number of edges, and hence can be applied to large networks, with millions of nodes.  ...  by kernel smoothing.  ... 
arXiv:1912.09068v1 fatcat:w76yl5vlo5enbl6ru3y3yj66hy

Automated smoother for the numerical decoupling of dynamics models

Marco Vilela, Carlos CH Borges, Susana Vinga, Ana Vasconcelos, Helena Santos, Eberhard O Voit, Jonas S Almeida
2007 BMC Bioinformatics  
Results: In this report we propose a robust, fully automated solution for signal extraction from time series, which is the prerequisite for the efficient reverse engineering of biological systems models  ...  Structure identification of dynamic models for complex biological systems is the cornerstone of their reverse engineering.  ...  S.Vinga also recognizes award PTDC/ EEA-ACR/69530/2006 "DynaMo -Dynamical modeling, control and optimization of metabolic networks" by Fundação para a Ciência e Tecnologia (FCT, Portugal).  ... 
doi:10.1186/1471-2105-8-305 pmid:17711581 pmcid:PMC2041957 fatcat:qxgleckp7jff5hi6buwo54dxyy
« Previous Showing results 1 — 15 out of 6,928 results