Acoustic keyword spotting in speech with applications to data mining

Keyword Spotting is the task of detecting keywords of interest within continu- ous speech. The applications of this technology range from call centre dialogue systems to covert speech surveillance devices. Keyword spotting is particularly well suited to data mining tasks such as real-time keyword monitoring and unre- stricted vocabulary audio document indexing. However, to date, many keyword spotting approaches have su®ered from poor detection rates, high false alarm rates, or slow execution times, thus reducing their commercial viability. This work investigates the application of keyword spotting to data mining tasks. The thesis makes a number of major contributions to the ¯eld of keyword spotting. The ¯rst major contribution is the development of a novel keyword veri¯cation method named Cohort Word Veri¯cation. This method combines high level lin- guistic information with cohort-based veri¯cation techniques to obtain dramatic improvements in veri¯cation performance, in particular for the problematic short duration target word class. The second major contribution is the development of a novel audio document indexing technique named Dynamic Match Lattice Spotting. This technique aug- ments lattice-based audio indexing principles with dynamic sequence matching techniques to provide robustness to erroneous lattice realisations. The resulting algorithm obtains signi¯cant improvement in detection rate over lattice-based audio document indexing while still maintaining extremely fast search speeds. The third major contribution is the study of multiple veri¯er fusion for the task of keyword veri¯cation. The reported experiments demonstrate that substantial improvements in veri¯cation performance can be obtained through the fusion of multiple keyword veri¯ers. The research focuses on combinations of speech background model based veri¯ers and cohort word veri¯ers. The ¯nal major contribution is a comprehensive study of the e®ects of limited training data for keyword spotting. This study is performed with consideration as to how these e®ects impact the immediate development and deployment of speech technologies for non-English languages.

Inferring neuronal network connectivity from spike data: A temporal data mining approach

Understanding the functioning of a neural system in terms of its underlying circuitry is an important problem in neuroscience. Recent d evelopments in electrophysiology and imaging allow one to simultaneously record activities of hundreds of neurons. Inferring the underlying neuronal connectivity patterns from such multi-neuronal spike train data streams is a challenging statistical and computational problem. This task involves finding significant temporal patterns from vast amounts of symbolic time series data. In this paper we show that the frequent episode mining methods from the field of temporal data mining can be very useful in this context. In the frequent episode discovery framework, the data is viewed as a sequence of events, each of which is characterized by an event type and its time of occurrence and episodes are certain types of temporal patterns in such data. Here we show that, using the set of discovered frequent episodes from multi-neuronal data, one can infer different types of connectivity patterns in the neural system that generated it. For this purpose, we introduce the notion of mining for frequent episodes under certain temporal constraints; the structure of these temporal constraints is motivated by the application. We present algorithms for discovering serial and parallel episodes under these temporal constraints. Through extensive simulation studies we demonstrate that these methods are useful for unearthing patterns of neuronal network connectivity.

Data mining approaches to software fault diagnosis

Automatic identification of software faults has enormous practical significance. This requires characterizing program execution behavior and the use of appropriate data mining techniques on the chosen representation. In this paper, we use the sequence of system calls to characterize program execution. The data mining tasks addressed are learning to map system call streams to fault labels and automatic identification of fault causes. Spectrum kernels and SVM are used for the former while latent semantic analysis is used for the latter The techniques are demonstrated for the intrusion dataset containing system call traces. The results show that kernel techniques are as accurate as the best available results but are faster by orders of magnitude. We also show that latent semantic indexing is capable of revealing fault-specific features.

Root cause diagnostics using temporal data mining

A method, system, and computer program product for fault data correlation in a diagnostic system are provided. The method includes receiving the fault data including a plurality of faults collected over a period of time, and identifying a plurality of episodes within the fault data, where each episode includes a sequence of the faults. The method further includes calculating a frequency of the episodes within the fault data, calculating a correlation confidence of the faults relative to the episodes as a function of the frequency of the episodes, and outputting a report of the faults with the correlation confidence.

Development of novel EST-SSR markers in common carp by data mining from public EST sequences

Expressed sequence tags (ESTs) are a source for microsatellite development. In the present study, EST-derived microsatelltes (EST-SSRs) were generated and characterized in the common carp (Cyprinus carpio) by data mining from updated public EST databases and by subsequent testing for polymorphism. About 5.5% (555) of 10,088 ESTs contain repeat motifs of various types and lengths with CA being the most abundant dinucleotide one. Out of the 60 EST-SSRs for which PCR primers were designed, 25 loci showed polymorphism in a common carp population with the alleles per locus ranging from 3 to 17 (mean 7). The observed (H-O) and expected (HE) heterozygosities of these EST-SSRs were 0.13-1.00 and 0.12-0.91, respectively. Six EST-SSR loci significantly deviated from the Hardy-Weinberg equilibrium (HWE) expectation, and the remaining 19 loci were in HWE. Of the 60 primer sets, the rates of polymorphic EST-SSRs were 42% in common carp, 17% in crucian carp (Carassius auratus), and 5% in silver carp (Hypophthalmichthys molitrix), respectively. These new EST-SSR markers would provide sufficient polymorphism for population genetic studies and genome mapping of the common carp and its closely related fishes. (c) 2007 Published by Elsevier B.V.

Confirmation of Data Mining Based Predictions of Protein Function

King, R. D. and Wise, P. H. and Clare, A. (2004) Confirmation of Data Mining Based Predictions of Protein Function. Bioinformatics 20(7), 1110-1118

An ontological characterization of time-series and state-sequences for data mining

Time-series and sequences are important patterns in data mining. Based on an ontology of time-elements, this paper presents a formal characterization of time-series and state-sequences, where a state denotes a collection of data whose validation is dependent on time. While a time-series is formalized as a vector of time-elements temporally ordered one after another, a state-sequence is denoted as a list of states correspondingly ordered by a time-series. In general, a time-series and a state-sequence can be incomplete in various ways. This leads to the distinction between complete and incomplete time-series, and between complete and incomplete state-sequences, which allows the expression of both absolute and relative temporal knowledge in data mining.

Nitrogen assimilation in Citrus based on CitEST data mining

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

New algorithms for automated phylogenetic reconstruction using artificial intelligence and data mining techniques

Academic and industrial research in the late 90s have brought about an exponential explosion of DNA sequence data. Automated expert systems are being created to help biologists to extract patterns, trends and links from this ever-deepening ocean of information. Two such systems aimed on retrieving and subsequently utilizing phylogenetically relevant information have been developed in this dissertation, the major objective of which was to automate the often difficult and confusing phylogenetic reconstruction process. ^ Popular phylogenetic reconstruction methods, such as distance-based methods, attempt to find an optimal tree topology (that reflects the relationships among related sequences and their evolutionary history) by searching through the topology space. Various compromises between the fast (but incomplete) and exhaustive (but computationally prohibitive) search heuristics have been suggested. An intelligent compromise algorithm that relies on a flexible “beam” search principle from the Artificial Intelligence domain and uses the pre-computed local topology reliability information to adjust the beam search space continuously is described in the second chapter of this dissertation. ^ However, sometimes even a (virtually) complete distance-based method is inferior to the significantly more elaborate (and computationally expensive) maximum likelihood (ML) method. In fact, depending on the nature of the sequence data in question either method might prove to be superior. Therefore, it is difficult (even for an expert) to tell a priori which phylogenetic reconstruction method—distance-based, ML or maybe maximum parsimony (MP)—should be chosen for any particular data set. ^ A number of factors, often hidden, influence the performance of a method. For example, it is generally understood that for a phylogenetically “difficult” data set more sophisticated methods (e.g., ML) tend to be more effective and thus should be chosen. However, it is the interplay of many factors that one needs to consider in order to avoid choosing an inferior method (potentially a costly mistake, both in terms of computational expenses and in terms of reconstruction accuracy.) ^ Chapter III of this dissertation details a phylogenetic reconstruction expert system that selects a superior proper method automatically. It uses a classifier (a Decision Tree-inducing algorithm) to map a new data set to the proper phylogenetic reconstruction method. ^

Data mining for lifetime prediction of metallic components

The ability to accurately predict the lifetime of building components is crucial to optimizing building design, material selection and scheduling of required maintenance. This paper discusses a number of possible data mining methods that can be applied to do the lifetime prediction of metallic components and how different sources of service life information could be integrated to form the basis of the lifetime prediction model