Explanation-Based Feature Construction

"Incorporating additional information from our prior domain knowledge can be the key to solving difficult classification tasks, especially when the available training data is limited. The crucial stage of feature construction, often done manually, plays a significant role in allowing such information to be incorporated into a learner. We propose algorithms for automated feature construction where available domain knowledge, even though imperfect and approximate, can be utilized by the learning system. Robustness is achieved by incorporating this prior knowledge in a task-specific manner, guided by the actual training examples. These goals are realized with Explanation-Based Learning (EBL). The EBL paradigm provides the necessary bridge between domain knowledge and the training examples, which allows us to design solutions that are conceptually well-formed and work for the right reason. The ideas of well-formed concepts and \working for the right reason"" are our guiding principles for supervised learning. Using these underlying principles, we propose three algorithms for incorporating prior domain knowledge into discriminative learning with different levels of interaction between the feature construction process and the final classifier learning. The first approach involves automated construction of generative models for phantom examples, which can be used to enhance the training data for subsequent classifier learning. Both the second and the third approaches involve the construction of semantic features. Each semantic feature encapsulates a well-formed concept which, ac- cording to the domain knowledge, corresponds to a conceptual difference between classes of objects. We illustrate and evaluate the proposed algorithms on the challenging problem of classifying offine handwritten Chinese characters, focusing on distinguishing difficult, mutually-similar pairs of characters. Empirical results show that our approaches can outperform the state-of-the-art algorithms."

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Computer Science

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2009

Author dc:creator

• Lim, Shiau Hong

Contributors dc:contributor

• DeJong, Gerald F.
• Roth, Dan
• Forsyth, David A.
• Amir, Eyal

• machine learning
• feature construction
• explanation-based learning
• pattern recognition
• handwriting recognition

Language dc:language

en

Identifier

Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Computer Science in the Graduate College of the University of Illinois at Urbana-Champaign, 2009

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/11753