Prepared by the LVQ Programming Team of the (2008)
Teuvo Kohonen, Jussi Hynninen, Jari Kangas, Jorma Laaksonen, Kari Torkkola
Speech processing comprises automatic speech recognition, speech synthesis, speech coding, speech enhancement, speaker recognition and verification, language identification, and so on. This section...
Sequential Feature Extraction Using Information-Theoretic Learning (2008)
Deniz Erdogmus, Kari Torkkola, Jose C. Principe
Abstract-- A classification system typically includes both a feature extractor and a classifier. The two components can be trained either sequentially or simultaneously. The former option has an...
Sensor Sequence Modeling for Driving (2008)
Kari Torkkola, Srihari Venkatesan, Huan Liu
Intelligent systems in automobiles need to be aware of the driving and driver context. Available sensor data stream has to be modeled and monitored in order to do so. Currently there exist no...
An approach to automate knowledge acquisition from natural signals, such as speech, is presented. The knowledge is extracted in the form of context-sensitive production rules that can be used to map...
WarpNet: Self-Organizing Time Warping (2007)
We describe "WarpNet", a time-warping algorithm for speech recognition based on neural nets. WarpNet is a self-organizing matching mechanism intended to replace dynamic programming...
List of abbreviations: HMM Hidden Markov model (2007)
Jyri Mantysalo, Kari Torkkolay, Teuvo Kohonen, Kari Torkkola
Mapping context dependent acoustic information into context independent form by LVQ
LVQ as a feature transformation for HMMs (2007)
Abstract. We present a new way to take advantage of the discriminative power of Learning Vector Quantization in combination with continuous density hidden Markov models. This is based on viewing LVQ...
We present experiments in using neural network based methods to initialize continuous observation density hidden Markov models (CDHMMs). Proper initialization provides an easy way to avoid excessive...
Copyright c fl1991-1995 by (2007)
Teuvo Kohonen, Jussi Hynninen, Jari Kangas, Jorma Laaksonen, Kari Torkkola
1 Contents
Sensor Selection for Maneuver Classification (2004)
Kari Torkkola, Srihari Venkatesan, Huan Liu
Abstract — To determine when to present information from various devices or services to the driver of an automobile, it is necessary to determine whether a driver is engaged in a difficult driving...
Data modelling in semiconductor industry is often challenged with complexity of massive mixed type datasets. Our main motivation is exploratory visualization of such data, usually for yield...
Feature extraction by non-parametric mutual information maximization (2003)
Kari Torkkola, Isabelle Guyon, André Elisseeff
We present a method for learning discriminative feature transforms using as criterion the mutual information between class labels and transformed features. Instead of a commonly used mutual...
Maximization Kari Torkkola, Kari Torkkola, Isabelle Guyon, Andre Elisseeff
We present a method for learning discriminative feature transforms using as criterion the mutual information between class labels and transformed features. Instead of a commonly used mutual...
Discriminative Features for Text Document Classification (2002)
The bag-of-words approach to text document representation typically results in vectors of the or-der of 5000 to 20000 components as the representation of documents. In order to make effective use of...
On feature extraction by mutual information maximization (2002)
In order to learn discriminative feature transforms, we discuss mutual information between class labels and transformed features as a criterion. Instead of Shannon’s definition we use measures...
Linear discriminant analysis in document classification (2001)
Document representation using the bag-of-words approach may require bringing the dimensionality of the representation down in order to be able to make effective use of various statistical...
Linear discriminant analysis in document classification (2001)
Document representation using the bag-of-words approach may require bringing the dimensionality of the representation down in order to be able to make effective use of various statistical...
Learning discriminative feature transforms to low dimensions in low dimensions (2001)
The marriage of Renyi entropy with Parzen density estimation has been shown to be a viable tool in learning discriminative feature transforms. However, it suffers from computational complexity...
Learning discriminative feature transforms to low dimensions in low dimensions (2001)
The marriage of Renyi entropy with Parzen density estimation has been shown to be a viable tool in learning discriminative feature transforms. However, it suffers from computational complexity...
Nonlinear feature transforms using maximum mutual information (2001)
Finding the right features is an essential part of a pattern recognition system. This can be accomplished either by selection or by a transform from a larger number of “raw” features. In this...
Visualizing class structure in data using mutual information (2000)
Abstract. We study linear dimension reducing transforms using maximum mutual information between transformed data and class labels as the criterion to learn the transforms. Renyi's quadratic...
Dimension reduction techniques for training polynomial networks (2000)
William M. Campbell, Kari Torkkola, Sreeram V. Balakrishnan
We propose two novel methods for reducing dimension in training polynomial networks. We consider the class of polynomial networks whose output is the weighted sum of a basis of monomials. Our first...
Mutual information in learning feature transformations (2000)
Kari Torkkola, William M. Campbell
We present feature transformations useful for exploratory data analysis or for pattern recognition. Transformations are learned from example data sets by maximizing the mutual information between...
Dimension reduction techniques for training polynomial networks (2000)
William M. Campbell, Kari Torkkola, Sreeram V. Balakrishnan
We propose two novel methods for reducing dimension in training polynomial networks. We consider the class of polynomial networks whose output is the weighted sum of a basis of monomials. Our first...
Mutual information in learning feature transformations (2000)
Kari Torkkola, William M. Campbell
We present feature transformations useful for exploratory data analysis or for pattern recognition. Transformations are learned from example data sets by maximizing the mutual information between...
Visualizing class structure in data using mutual information (2000)
Abstract. We study linear dimension reducing transforms using maximum mutual information between transformed data and class labels as the criterion to learn the transforms. Renyi’s quadratic...
Dimension reduction techniques for training polynomial networks (2000)
William M. Campbell, Kari Torkkola, Sreeram V. Balakrishnan
We propose two novel methods for reducing dimension in training polynomial networks. We consider the class of polynomial networks whose output is the weighted sum of a basis of monomials. Our first...
Evaluation Of Blind Signal Separation Methods (1999)
Daniel Schobben, Kari Torkkola, Paris Smaragdis
Recently, many new Blind Signal Separation (BSS) algorithms have been introduced. Authors evaluate the performance of their algorithms in various ways. Among these are speech recognition rates, plots...
Blind Separation For Audio Signals - Are We There Yet? (1999)
We attempt to give an overview of current research in blind separation of convolutive mixing of signals, concentrating on audio signals, and methods applicable thereof. We briefly enumerate some...
Blind Signal Separation In Communications: Making Use Of Known Signal Distributions (1998)
We apply maximum likelihood blind source separation [7] to complex valued signals mixed with complex valued nonstationary matrices. This case arises in radio communications with baseband signals. We...
Blind Separation of Radio Signals in Fading Channels (1997)
We apply information maximization / maximum likelihood blind source separation [2, 6] to complex valued signals mixed with complex valued nonstationary matrices. This case arises in radio...
Blind separation of delayed sources based on information maximization (1996)
Recently, Bell and Sejnowski have presented an approach to blind source separation based on the information maximization principle. We extend this approach into more general cases where the sources...
LVQ PAK: The Learning Vector Quantization Program Package (1996)
Teuvo Kohonen, Teuvo Kohonen, Teuvo Kohonen, Jussi Hynninen, Jussi Hynninen, Jussi Hynninen, ...
: Learning Vector Quantization (LVQ) is a group of algorithms applicable to statistical pattern recognition, in which the classes are described by a relatively small number of codebook vectors,...
Blind Separation Of Convolved Sources Based On Information Maximization (1996)
Blind separation of independent sources from their convolutive mixtures is a problem in many real world multi-sensor applications. In this paper we present a solution to this problem based on the...
Blind Separation Of Delayed Sources Based On Information Maximization (1996)
Recently, Bell and Sejnowski have presented an approach to blind source separation based on the information maximization principle. We extend this approach into more general cases where the sources...
Blind Separation Of Delayed Sources Based On Information Maximization (1996)
Recently, Bell and Sejnowski have presented an approach to blind source separation based on the information maximization principle. We extend this approach into more general cases where the sources...
New ways to use LVQ-codebooks together with hidden Markov models (1994)
We introduce a novel way to employ codebooks trained by Learning Vector Quantization together with hidden Markov models. In previous work, LVQ-codebooks have been used as frame labelers. The...
An efficient way to learn English grapheme-to-phoneme rules automatically (1993)
We present an efficient way to learn automatically grapheme-to-phoneme mapping rules for English by using Kohonen's concept of Dynamically Expanding Context. This method constructs rules that...
Teuvo Kohonen, Jari Kangas, Jorma Laaksonen, Kari Torkkola
Abstract. This paper is an overview of the program package LVQ PAK, which has been developed for convenient and effective application of Learning Vector Quantization algorithms. Two new features are...
Using SOMs as feature extractors for speech recognition (1992)
Jari Kangas, Kari Torkkola, Mikko Kokkonen
In this paper we demonstrate that the Self-Organizing Maps of Kohonen can be used as speech feature extractors that are able to take temporal context into account. We have investigated two...
We propose a novel initialization method for continuous observation density hidden Markov models (CDHMMs) that is based on Self-Organizing Maps (SOMs) and Learning Vector Quantization (LVQ) [6]. Our...
Short-Time feature vector based phonemic speech recognition with the aid of local context / (1991)
Thesis (doctoral)--Helsinki University of Technology, 1991.
Short-time feature vector based phonemic speech recognition with the aid of local context / (1991)
Nimiösivulla myös: Helsinki University of Technology, Faculty of Information Technology, Department of Computer Science, Laboratory of Computer and Information Science.
Improving Short-Time Speech Frame Recognition Results by Using Context (1991)
Kari Torkkola, Mikko Kokkonen, Mikko Kurimo, Pekka Utela
This paper focuses on comparing three approaches to improve the accuracy of classifying short-time speech frames into phoneme classes by taking into account the classifications of nearby frames, also...
Status Report Of The Finnish Phonetic Typewriter Project (1991)
Kari Torkkola, Jari Kangas, Pekka Utela, Sami Kaski, Mikko Kokkonen, Mikko Kurimo, ...
In connection to a speech recognizer, the aim of which is to produce phonemic transcriptions of arbitrary spoken utterances, we investigate the combined effect of several improvements at different...
Blind Separation Of Convolved Sources Based On Information Maximization
Kari Torkkola Motorola, Kari Torkkola
Blind separation of independent sources from their convolutive mixtures is a problem in many real world multi-sensor applications. In this paper we present a solution to this problem based on the...
Combining LVQ with continuous density hidden Markov models in speech recognition
We propose the use of Self-Organizing Maps (SOMs) and Learning Vector Quantization (LVQ) [5] as an initialization method for the training of the continuous observation density hidden Markov models...