On-line EM Algorithm for the Normalized Gaussian Network

By Masa-aki Sato, Shin Ishii. Neural Computation, 2000

311 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/606f7ac92d5170fab86e012c93fb95202d6a9823

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In General Health Professions

Citations in the corpus

Iterative RELIEF for Feature Weighting: Algorithms, Theories, and Applications (2007, background): “phasize the most recently observed data [26] by assuming”
Theory and Experiments on Vector Quantized Autoencoders (2018, methodology): “ This amounts to doing stochastic gradient in the space of both code-book embeddings and cluster assignments. These techniques have also been successfully used in minibatch K-means [29] and online EM [17, 28]. The generative process begins by sampling a sequence of discrete latent codes from an auto”
Distributed Asynchronous Online Learning for Natural Language Processing (2010, methodology): “Modern statistical NLP models are notoriously expensive to train, requiring the use of generalpurpose or specialized numerical optimization algorithms (e.g., gradient and coordinate ascent algorithms and variations on them like L-BFGS and EM) that iterate over training data many times.”
SkyAI: Highly modularized reinforcement learning library (2010, background): “config ={ Factor ={ resize(17, 5) [0]= ( 0, 0, 0, 0, 0) [1]= ( 1, 0, 0, 0, 0) [2]= ( 0, 0, 0, 0, 0) [3]= ( 0, 1, 0, 0, 0) [4]= ( 1, 0, 0, 0, 0) [5]= ( 0, 0, 0, 0, 0) [6]= ( 0, 1, 0, 0, 0) [7]= ( 0, 0, 0, 0, 0) [8]= ( 0, 0, 1, 0, 0) [9]= ( 0, 0, 0, 1, 0) [10]= ( 0, 0, 0, 0, 1) [11]= ( 0, 0, 0, 0, 0) ”
Online Learning for Latent Dirichlet Allocation (2010, methodology): “The problem is addressed by online coordinate ascent algorithms such as those described in [20] [21] [16] [17] [10].”
Scalable Learning of Collective Behavior (2012, background): “to handle extremely large data sets like online k-means [37], scalable k-means [38], and distributed k-means [39].”
Bayesian Computation Emerges in Generic Cortical Microcircuits through Spike-Timing-Dependent Plasticity (2013, result): “In contrast to previous rate-based models for probabilistic inference [28–30] every spike in our model has a clear semantic interpretation: one spike indicates the instantaneous assignment of a certain value to an abstract variable represented by the firing neuron.”
Online Model Selection Based on the Variational Bayes (2001, background): “Moreover, the EFH modelsincludea lotofinterestingmodels suchas normalized gaussiannetworks ( Sato & Ishii, 2000 ), hidden Markov models (Rabiner, 1989), mixture of gaussian models (Roberts, Husmeier, Rezek, & Penny, 1998), mixture of factor analyzers (Ghahramani & Beal, 2000), mixture of probabilist”
Opponent Modeling by Expectation–Maximization and Sequence Prediction in Simplified Poker (2017, unknown): “We use what Liang and Klein refer to as a stepwise EM algorithm [31], first proposed by Sato and Ishii [32], generalised by Cappé and Moulines [33], and applied to poker by Butterworth [34].”
On‐line expectation–maximization algorithm for latent data models (2007, methodology): “A speciﬁc instance of the proposed online EM algorithm has been derived by Sato and Ishii (2000) for maximum likelihood estimation in the so-called normalised Gaussian network; this algorithm was later extended by Sato (2000) to a canonical exponential family ( φ ( θ ) = θ in (11)) and a sketch of…”
Scalable learning of collective behavior based on sparse social dimensions (2009, unknown): “Then other k-means variants to handle extremely large data sets like On-line k-means [17], Scalable k-means [2], Incremental k-means [16] and distributed k-means [7] can be considered.”
Online Learning with Hidden Markov Models (2008, methodology): “where η(T) is a time-dependent discount factor used to remove the contribution to the sufficient statistics computed with earlier, inaccurate model parameters (e.g., Sato & Ishii, 2000 ); γlh(yT ; ˆ θ(T − 1)) is given by (see equation 2.10)”