Physical Sciences and Mathematics Commons^™

Using Symbolic Knowledge In The Umls To Disambiguate Words In Small Datasets With A Naive Bayes Classifier, Gondy Leroy, Thomas C. Rindflesch

CGU Faculty Publications and Research

Current approaches to word sense disambiguation use and combine various machine-learning techniques. Most refer to characteristics of the ambiguous word and surrounding words and are based on hundreds of examples. Unfortunately, developing large training sets is time-consuming. We investigate the use of symbolic knowledge to augment machine-learning techniques for small datasets. UMLS semantic types assigned to concepts found in the sentence and relationships between these semantic types form the knowledge base. A naïve Bayes classifier was trained for 15 words with 100 examples for each. The most frequent sense of a word served as the baseline. The effect of increasingly …

Reducing Redundancy In The Hypertree Decomposition Scheme, Peter Harvey, Aditya K. Ghose

Faculty of Informatics - Papers (Archive)

Hypertree decomposition is a powerful technique for transforming near-acyclic CSPs into acyclic CSPs. Acyclic CSPs have efficient, polynomial time solving techniques, and so these conversions are of interest to the constraints community. We present here an improvement on the opt-k-decomp algorithm for finding an optimal hypertree decomposition.

On Textured Image Analysis Using Wavelets., Mausumi Acharyya Dr.

Doctoral Theses

In image processing and computer vision research, we aim to derive better tools that give us different perspectives on the same image, allowing us to understand not only its content, but also its meaning and significance. Image processing can not compete with the human eye in terms of accuracy but it can outperform the latter easily on observational consistency, and ability to carry out detailed mathematical estimations. With time, image processing research has broadened from the basic pixel-based low- level operations to high-level analysis, that now includes the use of artificially intelligent techniques for image interpretation and understanding. These new …

Machine Learning Approaches For Determining Effective Seeds For K -Means Algorithm, Kaveephong Lertwachara

Doctoral Dissertations

In this study, I investigate and conduct an experiment on two-stage clustering procedures, hybrid models in simulated environments where conditions such as collinearity problems and cluster structures are controlled, and in real-life problems where conditions are not controlled. The first hybrid model (NK) is an integration between a neural network (NN) and the k-means algorithm (KM) where NN screens seeds and passes them to KM. The second hybrid (GK) uses a genetic algorithm (GA) instead of the neural network. Both NN and GA used in this study are in their simplest-possible forms.

In the simulated data sets, I investigate two …

Certain Pattern Recognition Tasks For Data Mining Problems., Pabitra Mitra Dr.

Doctoral Theses

Pattern recognition (PR) is an activity that we humans normally excel in. We do it almost all the time, and without conscious effort. We receive information via our various sensory organs, which is processed instantaneously by our brain so that, almost immediately, we are able to identify the source of the information, without having made any perceptible effort. What is even more impressive is the accuracy with which we can perform recognition tasks even under non-ideal conditions, for instance, when the information that needs to be processed is vague, imprecise or even incomplete. In fact, most of our day-to-day activities …

Rapid Development Of Hindi Named Entity Recognition Using Conditional Random Fields And Feature Induction, Wei Li, Andrew Mccallum

Andrew McCallum

This paper describes our application of Conditional Random Fields (CRFs) with feature induction to a Hindi named entity recognition task. With only five days development time and little knowledge of this language, we automatically discover relevant features by providing a large array of lexical tests and using feature induction to automatically construct the features that most increase conditional likelihood. In an effort to reduce overfitting, we use a combination of a Gaussian prior and early-stopping based on the results of 10-fold cross validation.

Modular Machine Learning Methods For Computer-Aided Diagnosis Of Breast Cancer, Mia Kathleen Markey '94

Doctoral Dissertations

The purpose of this study was to improve breast cancer diagnosis by reducing the number of benign biopsies performed. To this end, we investigated modular and ensemble systems of machine learning methods for computer-aided diagnosis (CAD) of breast cancer. A modular system partitions the input space into smaller domains, each of which is handled by a local model. An ensemble system uses multiple models for the same cases and combines the models' predictions.

Five supervised machine learning techniques (LDA, SVM, BP-ANN, CBR, CART) were trained to predict the biopsy outcome from mammographic findings (BIRADS™) and patient age based on a …

Smoke And Mirrors Or Science? Teaching Law With Computers - A Reply To Cass Sunstein On Artificial Intelligence And Legal Science, Eric A. Engle

Eric A. Engle

The article explores the possibilities and limits of AI for teaching and modeling law.

Using Boosting To Simplify Classification Models, V. Wheway

Faculty of Informatics - Papers (Archive)

Ensemble classification techniques such as bagging, boosting and arcing algorithms have been shown to lead to reduced classification errors on unseen cases and seem immune to the problem of overfitting. Several explanations for the reduction in generalisation error have been presented, with recent authors defining and applying diagnostics such as "edge" and "margin". These measures provide insight into the behaviour of ensemble classifiers, but can they be exploited further? In this paper, a four-stage classification procedure in introduced, which is based on an extension of edge and margin analysis. This new procedure allows inverse sub-contexts and difficult border regions to …

On Some Self-Organizing Models And Their Applications., Amitava Dutta Dr.

Doctoral Theses

Abstract: Self-organizing neural network models constitute the main theme of this thesis. Some well-known self-organizing models are surveyed and their properties are discussed. The application areas on which the thesis focuses are briefly described.This thesis deals with Artificial Neural Network models, in particular, Self- organizing (unsupervisnd) models. We develop here a few self-organizing neural net- work models to solve certain problems which are well studied in the areas of Image Processing and Computationel Geometry and have wide applications in shape eztrac- tion and optimization.1.1 Artificial neural networkThe study of Biological Neural Networks originally comes under biological sciences. They deal with …

The Application And Performance Of A Generic Task Routine Decision Making Algorithm To Recipe Selection In Meal Planning, Michelle M. Cox

Theses and Dissertations - UTB/UTPA

A nutritional meal planning system was implemented to test the effectiveness of a previously developed routine decision making algorithm. The combinatorics involved in ordering recipes in all possible combinations to produce variability in a meal plan and provide sufficient nutrition is conceptually intensive. Meal planning involves selection of food to eat to fulfill a person's nutritional and personal preferences. This thesis demonstrates meal planning as a decision making problem and demonstrates the utility of the routine decision making algorithm by solving this problem. Generic Tasks, identified through artificial intelligence research, provides the basis for this algorithm. It uses user preferences …

Feature Evaluation, Classification And Rule Generation Using Fuzzy Sets And Neural Networks., Rajat Kumar De Dr.

Doctoral Theses

Pattern recognition and machine learning form a major area of research and develop- ment activity that encompasses the processing of pictorial and other non-numerical information obtained from the interaction between science, technology and society. A motivation for the spurt of activity in this field is the need for people to com- municate with the computing machines in their natural mode of communication. Another important motivation is that the scientists are also concerned with the idea of designing and making intelligent machines that can carry out certain tasks that we human beings do. The most salient outcome of these is the …

Cataloging Expert Systems: Optimism And Frustrated Reality, William Olmstadt

E-JASL 1999-2009 (Volumes 1-10)

There is little question that computers have profoundly changed how information professionals work. The process of cataloging and classifying library materials was one of the first activities transformed by information technology. The introduction of the MARC format in the 1960s and the creation of national bibliographic utilities in the 1970s had a lasting impact on cataloging. In the 1980s, the affordability of microcomputers made the computer accessible for cataloging, even to small libraries. This trend toward automating library processes with computers parallels a broader societal interest in the use of computers to organize and store information. Following World War II, …

Designing Electronic Casebooks That Talk Back: The Cato Program, Kevin D. Ashley

Articles

Electronic casebooks offer important benefits of flexibility in control of presentation, connectivity, and interactivity. These additional degrees of freedom, however, also threaten to overwhelm students. If casebook authors and instructors are to achieve their pedagogical goals, they will need new methods for guiding students. This paper presents three such methods developed in an intelligent tutoring environment for engaging students in legal role-playing, making abstract concepts explicit and manipulable, and supporting pedagogical dialogues. This environment is built around a program known as CATO, which employs artificial intelligence techniques to teach first-year law students how to make basic legal arguments with cases. …

Life And Evolution In Computers, Melanie Mitchell

Computer Science Faculty Publications and Presentations

This paper argues for the possibility of 'artificial life' and computational evolution, first by discussing (via a highly simplified version) John von Neumann's self-reproducing automaton and then by presenting some recent work focusing on computational evolution, in which 'cellular automata', a form of parallel and decentralized computing system, are evolved via 'genetic algorithms'. It is argued that such in silico experiments can help to make sense of the question of whether we can eventually build computers that are intelligent and alive.

Pattern Classification Using Genetic Algorithms., Sanghamitra Bandyopadhyay Dr.

Doctoral Theses

Pattern recognition and machine learning form a major area of research and develop- ment activity that encompasses the processing of pictorial and other non-numerical information obtained from the interaction between science, technology and society. A motivation for the spurt of activity in this field is the need for people to com- municate with the computing machines in their natural mode of communication. Another important motivation is that the scientists are also concerned with the idea of designing and making intelligent machines that can carry out certain tasks that we human beings do. The most salient outcome of these is the …

The Round Table Model: A Web-Oriented, Agent-Based Approach To Decision-Support Applications, Kym J. Pohl, Jens G. Pohl

Collaborative Agent Design (CAD) Research Center

Not unlike King Arthur relying on the infamous Round Table as the setting for consultation with his most trusted experts, agent-based, decision-support systems provide human decision makers with a means of solving complex problems through collaboration with collections of both human and computer-based expert agents. The Round Table Framework provides a formalized architecture together with a set of development and execution tools which can be utilized to design, develop, and execute agent-based, decision-support applications. Based on a three-tier architecture, Round Table incorporates forefront technologies including distributed-object servers, inference engines, and web-based presentation to provide a framework for collaborative, agent-based decision …

On The Developement Of An Optical Character Recognition(Ocr) System For Printed Bangla Script., Umapada Pal Dr.

Doctoral Theses

This thesis concerns OCR development of machine printed text in an Indian lan- guage, Bangla (Bengali) which is the fourthmost popular language in the world and the secondmost popular language in India.1.1 Optical Character Recognition Optical Character Recognition (OCR) is a process of automatic computer recog- nition of characters in optically scanned and digitized pages of text. OCR is ene of the most fascinating and challenging areas of pattern recognition with various practical applications. It can contribute tremendously to the advancement of an automation process and can improve the interface between man and machine in many applications, including office automation …

Neuro Fuzzy Reasoning For Pattern Classification And Object Recognition., Jayati Ghosh Dr.

Doctoral Theses

In real world, pattern classification and object recognition problems are faced with fuzzi- ness that is connected with diverse facets of cognitive activity of the human being. An origin of sources of fuzziness is related to labels expressed in feature space as well as to labels of classes taken into account in classification and /or recognition procedures. Though a lot of scientific efforts have already been dedicated to pattern recognition problems, especially to classification procedures, still pattern recognition is confronted with a continuous challenge coming from a human being who can perform lot of ex- tremely complex classification tasks by …

Design And Design Centers In Engineering Education, Clive L. Dym

All HMC Faculty Publications and Research

This paper is intended to be the opening salvo of the workshop, Computing Futures in Engineering Design (Dym, 1997). Thus, I want to take this privileged moment to ask you to think with me about the role of design in engineering. In particular, I want to reflect upon how design is articulated and how design is taught; about the role of design in engineering education and in the practice of engineering; and about the role that could be played locally and, perhaps, nationally by a center devoted to design education. Because I teach here at Harvey Mudd College (HMC), …

A Complex-Systems Perspective On The “Computation Vs. Dynamics” Debate In Cognitive Science, Melanie Mitchell

Computer Science Faculty Publications and Presentations

I review the purported opposition between computational and dynamical approaches in cognitive science. I argue that both computational and dynamical notions will be necessary for a full explanatory account of cognition, and give a perspective on how recent research in complex systems can lead to a much needed rapprochement between computational and dynamical styles of explanation.

A Methodology For The Selection Of A Paradigm Of Reasoning Under Uncertainty In Expert System Development, Vivian Campbell

Theses: Doctorates and Masters

The aim of this thesis is to develop a methodology for the selection of a paradigm of reasoning under uncertainty for the expert system developer. This is important since practical information on how to select a paradigm of reasoning under uncertainty is not generally available. The thesis explores the role of uncertainty in an expert system and considers the process of reasoning under uncertainty. The possible sources of uncertainty are investigated and prove to be crucial to some aspects of the methodology. A variety of Uncertainty Management Techniques (UMTs) are considered, including numeric, symbolic and hybrid methods. Considerably more information …

Learning To See Analogies: A Connectionist Exploration, Douglas S. Blank

Computer Science Faculty Research and Scholarship

The goal of this dissertation is to integrate learning and analogy-making. Although learning and analogy-making both have long histories as active areas of research in cognitive science, not enough attention has been given to the ways in which they may interact. To that end, this project focuses on developing a computer program, called Analogator, that learns to make analogies by seeing examples of many different analogy problems and their solutions. That is, it learns to make analogies by analogy. This approach stands in contrast to most existing computational models of analogy in which particular analogical mechanisms are assumed a priori …

Simulation Study Of Learning Automata Games In Automated Highway Systems, Cem Unsal, Pushkin Kachroo, John S. Bay

Electrical & Computer Engineering Faculty Research

One of the most important issues in Automated Highway System (AHS) deployment is intelligent vehicle control. While the technology to safely maneuver vehicles exists, the problem of making intelligent decisions to improve a single vehicle’s travel time and safety while optimizing the overall traffic flow is still a stumbling block. We propose an artificial intelligence technique called stochastic learning automata to design an intelligent vehicle path controller. Using the information obtained by on-board sensors and local communication modules, two automata are capable of learning the best possible (lateral and longitudinal) actions to avoid collisions. This learning method is capable of …

Development Of A Standard Set Of Indicators And Metrics For Artificial Intelligence (Ai) And Expert System (Es) Software Development Efforts, Derek F. Cossey

Theses and Dissertations

The purpose of this research was to identify a standard set of indicators and metrics that can be used by program managers to improve their abilities to direct development efforts involving Artificial Intelligence (AI) and Expert Systems (ES). This research addresses two objectives. The first objective is to identify an appropriate set of software indicators and metrics to be used by government program offices for the management of development efforts involving software systems for AI and ES. The second objective is to demonstrate how the resources of the National Software Data and Information Repository (NSDIR) can be used in order …

Neuro-Fuzzy Models For Classification And Rule Generation., Sushmita Mitra Dr.

Doctoral Theses

Machine recognition [1, 2] of patterns can be viewed as a two-fold task, consisting of learning the invariant and common properties of a set of samples characterizing a class, and of deciding a new sample as a possible member of the class by noting that it has properties common to those of the set of samples. In other words, pattern recognition by computers can be described as a transformation from the measurenment space M to the feature space F and finally to the decision space D (1), i.e., M ⟶F⟶D.Here, the mapping 6 : F⟶D is the decision function and …

Connectionist Models For Certain Tasks Related To Object Recognition., Jayanta Basak Dr.

Doctoral Theses

Recognition of objects in an image, according to Suetens et al. [1), relers to the task of finding and labeling parts of a two-dimensional image of a scene that correspond to the real objects in the scene. Object recognition is necessary in a variety of domains like robot navigation, aerial imagery analysis, industrial inspection and so on. Normally, different strategies for object recognition (1-(5] involve establishing some model for each object, i.e., some general description of each object, and then labeling different parts of the scene according to the knowledge about the models.Object models can have two-dimensional (2D) or three-climensional …

An Integrated Framework For Learning And Reasoning, Christophe G. Giraud-Carrier, Tony R. Martinez

Faculty Publications

Learning and reasoning are both aspects of what is considered to be intelligence. Their studies within AI have been separated historically, learning being the topic of machine learning and neural networks, and reasoning falling under classical (or symbolic) AI. However, learning and reasoning are in many ways interdependent. This paper discusses the nature of some of these interdependencies and proposes a general framework called FLARE, that combines inductive learning using prior knowledge together with reasoning in a propositional setting. Several examples that test the framework are presented, including classical induction, many important reasoning protocols and two simple expert systems.

Interview: Brenda Laurel, Jason Challas

SWITCH

This interview with Brenda Laurel, Virtual Reality (VR) author and thinker, discusses the applications and challenges of VR. Creating an emphatic experience using VR technology is possible, but the challenge lies in designing an environment that models the senses to stimulate emotions. VR enables experiences of different genders, but physiological differences between the sexes exist and are important to understand. However, technology used to create the environment and simulation of physical objects in VR is only in the developmental stage. Laurel believes in the importance of keeping the mind grounded in the physical body, in order to strengthen the appreciation …

On Lexical And Syntactic Processing Of Bangla Language By Computer., Probal Sengupta Dr.

Doctoral Theses

A distinctive intelligent trait of human beings is the ability to carry out meaningful communication through language. The communication may be direct as in spoken conversation or indirect as in written form, through the audio-visual media, etc. Linguistic ability in humans have fascinated scholars ever since man first learnt to use language. Linguistics, the branch of study involved in studying the nature of human linguistic communication, is perhaps as old as language itself. The invention of the computer added a new dimension to linguistics. Making the computer emu- late human linguistic behaviour was taken up as a challenge by computer …