Computer Sciences Commons^™

On-Line Bayesian Speaker Adaptation By Using Tree-Structured Transformation And Robust Priors, Shaojun Wang, Yunxin Zhao

Kno.e.sis Publications

This paper presents new results by using our previously proposed on-line Bayesian learning approach for affine transformation parameter estimation in speaker adaptation. The on-line Bayesian learning technique allows updating parameter estimates after each utterance and it can accommodate flexible forms of transformation functions as well as prior probability density functions. We show through experimental results the robustness of heavy tailed priors to mismatch in prior density estimation. We also show that by properly choosing the transformation matrices and depths of hierarchical trees, recognition performance improved significantly.

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 16, Number 6, June 2000, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

A fourteen page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

Local Properties Of Query Languages, Guozhu Dong, Leonid Libkin, Limsoon Wong

Kno.e.sis Publications

In this paper we study the expressiveness of local queries. By locality we mean — informally — that in order to check if a tuple belongs to the result of a query, one only has to look at a certain predetermined portion of the input. Examples include all relational calculus queries. We start by proving a general result describing outputs of local queries. This result leads to many easy inexpressibility proofs for local queries. We then consider a closely related property, namely, the bounded degree property. It describes the outputs of local queries on structures that locally look “simple.” Every …

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 16, Number 5, May 2000, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

A nineteen page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 16, Number 4, April 2000, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

An eight page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 16, Number 3, March 2000, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

A ten page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 16, Number 2, February 2000, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

A ten page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

The Space Of Jumping Emerging Patterns And Its Incremental Maintenance, Jinyan Li, Kotagiri Ramamohanarao, Guozhu Dong

Kno.e.sis Publications

The concept of jumping emerging patterns (JEPs) has been proposed to describe those discriminating features which only occur in the positive training instances but do not occur in the negative class at all; JEPs have been used to construct classifiers which generally provide better accuracy than the state-of-the-art classifiers such as C4.5. The algorithms for maintaining the space of jumping emerging patterns (JEP space) are presented in this paper. We prove that JEP spaces satisfy the property of convexity. Therefore JEP spaces can be concisely represented by two bounds: consisting respectively of the most general elements and the most specific …

Separating Auxiliary Arity Hierarchy Of First-Order Incremental Evaluation Using (3+1)-Ary Input Relations, Guozhu Dong, Louxin Zhang

Kno.e.sis Publications

Presents a first-order incremental evaluation system that uses first-order queries to maintain a database view defined by a non-first-order query. Reduction of the arity of queries to understand the power of foies; Use of a key lemma for proving a query which encodes the multiple parity problem.

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 16, Number 1, January 2000, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

An eight page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

Der Kontraktionssatz Auf Metrischen Raumen Und Verallgemeinerungen, Pascal Hitzler

Computer Science and Engineering Faculty Publications

No abstract provided.

Imprecise Answers In Distributed Environments: Estimation Of Information Loss For Multi-Ontology Based Query Processing, Eduardo Mena, Vipul Kashyap, Arantza Illarramendi, Amit P. Sheth

Kno.e.sis Publications

The World Wide Web is fast becoming a ubiquitous computing environment. Prevalent keyword-based search techniques are scalable, but are incapable of accessing information based on concepts. We investigate the use of concepts from multiple, real-world pre-existing, domain ontologies to describe the underlying data content and support information access at a higher level of abstraction. It is not practical to have a single domain ontology to describe the vast amounts of data on the Web. In fact, we expect multiple ontologies to be used as different world views and present an approach to "browse" ontologies as a paradigm for information access. …

Optimization Techniques For Data Intensive Decision Flows, Richard Hull, Francois Llirbat, Bharat Kumar, Gang Zhou, Guozhu Dong, Jianwen Su

Kno.e.sis Publications

For an enterprise to take advantage of the opportunities afforded by electronic commerce it must be able to make decisions about business transactions in near-real-time. In the coming era of segment-of-one marketing, these decisions will be quite intricate, so that customer treatments can be highly personalized, reflecting customer preferences, the customer's history with the enterprise, and targeted business objectives. This paper describes a paradigm called “decision flows” for specifying a form of incremental decision-making that can combine diverse business factors in near-real-time.

This paper introduces and empirically analyzes a variety of optimization strategies for decision flows that are “data-intensive”, i.e. …

Characterizations Of Classes Of Programs By Three-Valued Operators, Anthony K. Seda, Pascal Hitzler

Computer Science and Engineering Faculty Publications

Several important classes of normal logic programs, including the classes of acyclic, acceptable, and locally hierarchical programs, have the property that every program in the class has a unique two-valued supported model. In this paper, we call such classes unique supported model classes. We analyse and characterize these classes by means of operators on three-valued logics. Our studies will motivate the definition of a larger unique supported model class which we call the class of Phi-accessible programs. Finally, we show that the class of Phi -accessible programs is computationally adequate in that every partial recursive function can be implemented by …

On-Line Bayesian Tree-Structured Transformation Of Hidden Markov Models For Speaker Adaptation, Shaojun Wang, Yunxin Zhao

Kno.e.sis Publications

This paper presents a new recursive Bayesian learning approach for transformation parameter estimation in speaker adaptation. Our goal is to incrementally transform (or adapt) the entire set of HMM parameters for a new speaker or new acoustic enviroment from a small amount of adaptation data. By establishing a clustering tree of HMM Gaussian mixture components, the finest affine transformation parameters for individual HMM Gaussian mixture components can be dynamically searched. The on-line Bayesian learning technique proposed in our recent work is used for recursive maximum a posteriori estimation of affine transformation parameters. Speaker adaptation experiments using a 26-letter English alphabet …

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 15, Number 9, November 1999, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

A fourteen page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 15, Number 8, October 1999, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

A twelve page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

Critical It Role In Healthcare: National Agenda, Applications, And Technologies, Jack Corley, Warren Karp, Amit P. Sheth

Kno.e.sis Publications

We share a widely-held vision that tomorrow's health care will be delivered by a distributed team using a knowledge based process that is focused on prevention and wellness. Caregivers will be able to treat and monitor patients where and when needed, capturing data as a natural byproduct of care delivery. Clinical teams with a wide range of skills and expertise will deliver consistent, quality care based on timely, situation-specific knowledge and guidance derived from outcomes evidence. That same knowledge will support advances in healthcare research and education. Home care and remote medicine will increase, reducing the need for high-cost hospitalization. …

Efficient Support For Decision Flows In E-Commerce Applications, Richard Hull, Francois Llirbat, Jianwen Su, Guozhu Dong, Bharat Kumar, Gang Zhou

Kno.e.sis Publications

In the coming era of segment-of-one marketing, decisions about business transactions will be quite intricate, so that customer treatments can be highly individualized, reflecting customer preferences, targeted business objectives, etc. This paper describes a paradigm called “decision flows” for specifying a form of incremental decision-making that can combine a myriad of diverse business factors and be executed in near-realtime. Starting with initial input, a decision flow will iteratively gather and derive additional information until a conclusion is reached. Decision flows can be specified in a rules-based manner that generalizes so-called “business rules” but provides more structure than traditional expert systems. …

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 15, Number 7, September 1999, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

A ten page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

A Multilevel Secure Workflow Management System, Myong H. Kang, Judith N. Froscher, Amit P. Sheth, Krzysztof J. Kochut, John A. Miller

Kno.e.sis Publications

The Department of Defense (DoD) needs multilevel secure (MLS) workflow management systems to enable globally distributed users and applications to cooperate across classification levels to achieve mission critical goals. An MLS workflow management system that allows a user to program multilevel mission logic, to securely coordinate widely distributed tasks, and to monitor the progress of the workflow across classification levels is required. In this paper, we present a roadmap for implementing MLS workflows and focus on a workflow builder that is a graphical design tool for specifying such workflows.

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 15, Number 6, June 1999, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

An eight page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 15, Number 5, May 1999, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

An eight page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

How Proteins Fold, Flex, And Bind Other Molecules, Leslie A. Kuhn, Volker Schnecke, Michael L. Raymer, Paul C. Sanschagrin

Kno.e.sis Publications

No abstract provided.

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 15, Number 4, April 1999, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

An eight page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 15, Number 3, March 1999, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

A ten page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

Data Integration By Describing Sources With Constraint Databases, Xun Cheng, Guozhu Dong, Tzekwan Lau, Jianwen Su

Kno.e.sis Publications

We develop a data integration approach for the efficient evaluation of queries over autonomous source databases. The approach is based on some novel applications and extensions of constraint database techniques. We assume the existence of a global database schema. The contents of each data source are described using a set of constraint tuples over the global schema; each such tuple indicates possible contributions from the source. The “source description catalog” (SDC) of a global relation consists of its associated constraint tuples. Such a method of description is advantageous since it is flexible to add new sources and to modify existing …

Efficient Mining Of Partial Periodic Patterns In Time Series Database, Jiawei Han, Guozhu Dong, Yiwen Yin

Kno.e.sis Publications

Partial periodicity search, i.e., search for partial periodic patterns in time-series databases, is an interesting data mining problem. Previous studies on periodicity search mainly consider finding full periodic patterns, where every point in time contributes (precisely or approximately) to the periodicity. However, partial periodicity is very common in practice since it is more likely that only some of the time episodes may exhibit periodic patterns.

We present several algorithms for efficient mining of partial periodic patterns, by exploring some interesting properties related to partial periodicity, such as the Apriori property and the max-subpattern hit set property, and by shared mining …

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 15, Number 2, February 1999, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

A ten page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

Multivalued Mappings, Fixed-Point Theorems And Disjunctive Databases, Pascal Hitzler, Anthony K. Seda

Computer Science and Engineering Faculty Publications

In this paper, we discuss the semantics of disjunctive programs and databases and show how multivalued mappings and their fixed points arise naturally within this context. A number of fixed-point theorems for multivalued mappings are considered, some of which are already known and some of which are new. The notion of a normal derivative of a disjunctive program is introduced. Normal derivatives are normal logic programs which are determined by the disjunctive program. Thus, the well-known single-step operator associated with a normal derivative is single-valued, and its fixed points can be found by well-established means. It is shown how fixed …