Results 1 
5 of
5
Solving multiclass learning problems via errorcorrecting output codes
 Journal of Artificial Intelligence Research
, 1995
"... Multiclass learning problems involve nding a de nition for an unknown function f(x) whose range is a discrete set containing k>2values (i.e., k \classes"). The de nition is acquired by studying collections of training examples of the form hx i;f(x i)i. Existing approaches to multiclass learning ..."
Abstract

Cited by 564 (9 self)
 Add to MetaCart
Multiclass learning problems involve nding a de nition for an unknown function f(x) whose range is a discrete set containing k>2values (i.e., k \classes"). The de nition is acquired by studying collections of training examples of the form hx i;f(x i)i. Existing approaches to multiclass learning problems include direct application of multiclass algorithms such as the decisiontree algorithms C4.5 and CART, application of binary concept learning algorithms to learn individual binary functions for each of the k classes, and application of binary concept learning algorithms with distributed output representations. This paper compares these three approaches to a new technique in which errorcorrecting codes are employed as a distributed output representation. We show that these output representations improve the generalization performance of both C4.5 and backpropagation on a wide range of multiclass learning tasks. We also demonstrate that this approach is robust with respect to changes in the size of the training sample, the assignment of distributed representations to particular classes, and the application of over tting avoidance techniques such as decisiontree pruning. Finally,we show thatlike the other methodsthe errorcorrecting code technique can provide reliable class probability estimates. Taken together, these results demonstrate that errorcorrecting output codes provide a generalpurpose method for improving the performance of inductive learning programs on multiclass problems. 1.
ErrorCorrecting Output Codes: A General Method for Improving Multiclass Inductive Learning Programs
 IN PROCEEDINGS OF AAAI91
, 1991
"... Multiclass learning problems involve finding a definition for an unknown function f(x) whose range is a discrete set containing k ? 2 values (i.e., k "classes"). The definition is acquired by studying large collections of training examples of the form hx i ; f(x i )i. Existing approaches to this pro ..."
Abstract

Cited by 88 (7 self)
 Add to MetaCart
Multiclass learning problems involve finding a definition for an unknown function f(x) whose range is a discrete set containing k ? 2 values (i.e., k "classes"). The definition is acquired by studying large collections of training examples of the form hx i ; f(x i )i. Existing approaches to this problem include (a) direct application of multiclass algorithms such as the decisiontree algorithms ID3 and CART, (b) application of binary concept learning algorithms to learn individual binary functions for each of the k classes, and (c) application of binary concept learning algorithms with distributed output codes such as those employed by Sejnowski and Rosenberg in the NETtalk system. This paper compares these three approaches to a new technique in which BCH errorcorrecting codes are employed as a distributed output representation. We show that these output representations improve the performance of ID3 on the NETtalk task and of backpropagation on an isolatedletter speechrecognition t...
Achieving HighAccuracy TexttoSpeech with Machine Learning
 In Data mining in speech synthesis
, 1997
"... In 1987, Sejnowski and Rosenberg developed their famous NETtalk system for English textto speech. This chapter describes a machine learning approach to texttospeech that builds upon and extends the initial NETtalk work. Among the many extensions to the NETtalk system were the following: a differe ..."
Abstract

Cited by 24 (2 self)
 Add to MetaCart
In 1987, Sejnowski and Rosenberg developed their famous NETtalk system for English textto speech. This chapter describes a machine learning approach to texttospeech that builds upon and extends the initial NETtalk work. Among the many extensions to the NETtalk system were the following: a different learning algorithm, a wider input "window", errorcorrecting output coding, a righttoleft scan of the word to be pronounced (with the results of each decision influencing subsequent decisions), and the addition of several useful input features. These changes yielded a system that performs much better than the original NETtalk system. After training on 19,002 words, the system achieves 93.7% correct pronunciation of individual phonemes and 64.8% correct pronunciation of whole words (where the pronunciation must exactly match the dictionary pronunciation to be correct). Based on the judgements of three human participants in a blind assessment study, our system was estimated to have a seri...
ErrorCorrection on NonStandard Communication Channels
, 2003
"... Many communication systems are poorly modelled by the standard channels assumed in the information theory literature, such as the binary symmetric channel or the additive white Gaussian noise channel. Real systems suffer from additional problems including timevarying noise, crosstalk, synchronizat ..."
Abstract

Cited by 3 (0 self)
 Add to MetaCart
Many communication systems are poorly modelled by the standard channels assumed in the information theory literature, such as the binary symmetric channel or the additive white Gaussian noise channel. Real systems suffer from additional problems including timevarying noise, crosstalk, synchronization errors and latency constraints. In this thesis, lowdensity paritycheck codes and codes related to them are applied to nonstandard channels. First, we look
Construction of quasicyclic codes
"... The class of QuasiCyclic Error Correcting Codes is investigated. It is shown that they contain many of the best known binary and nonbinary codes. Tables of rate 1/p and (p − 1)/p QuasiCyclic (QC) codes are constructed, which are a compilation of previously best known codes as well as many new code ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
The class of QuasiCyclic Error Correcting Codes is investigated. It is shown that they contain many of the best known binary and nonbinary codes. Tables of rate 1/p and (p − 1)/p QuasiCyclic (QC) codes are constructed, which are a compilation of previously best known codes as well as many new codes constructed using exhaustive, and other more sophisticated search techniques. Many of these binary codes attain the known bounds on the maximum possible minimum distance, and 13 improve the bounds. The minimum distances and generator polynomials of all known best codes are given. The search methods are outlined and the weight divisibility of the codes is noted. The weight distributions of some sth Power Residue (PR) codes and related rate 1/s QC codes are found using the link established between PR codes and QC codes. Subcodes of the PR codes are found by deleting certain circulant matrices in the corresponding QC code. They are used as a starting