Date on Master's Thesis/Doctoral Dissertation

12-2015

Document Type

Doctoral Dissertation

Degree Name

Ph. D.

Department

Computer Engineering and Computer Science

Degree Program

Computer Science and Engineering, PhD

Committee Chair

Frigui, Hichem

Committee Co-Chair (if applicable)

Nasraoui, Olfa

Committee Member

Nasraoui, Olfa

Committee Member

Zurada, Jacek

Committee Member

Lauf, Adrian

Committee Member

Hardin, Tim

Author's Keywords

multiple instance learning; fuzzy inference; ANFIS; mamdani; MI-ANFIS; MI-Mamdani

Abstract

A novel fuzzy learning framework that employs fuzzy inference to solve the problem of multiple instance learning (MIL) is presented. The framework introduces a new class of fuzzy inference systems called Multiple Instance Fuzzy Inference Systems (MI-FIS). Fuzzy inference is a powerful modeling framework that can handle computing with knowledge uncertainty and measurement imprecision effectively. Fuzzy Inference performs a non-linear mapping from an input space to an output space by deriving conclusions from a set of fuzzy if-then rules and known facts. Rules can be identified from expert knowledge, or learned from data. In multiple instance problems, the training data is ambiguously labeled. Instances are grouped into bags, labels of bags are known but not those of individual instances. MIL deals with learning a classifier at the bag level. Over the years, many solutions to this problem have been proposed. However, no MIL formulation employing fuzzy inference exists in the literature. In this dissertation, we introduce multiple instance fuzzy logic that enables fuzzy reasoning with bags of instances. Accordingly, different multiple instance fuzzy inference styles are proposed. The Multiple Instance Mamdani style fuzzy inference (MI-Mamdani) extends the standard Mamdani style inference to compute with multiple instances. The Multiple Instance Sugeno style fuzzy inference (MI-Sugeno) is an extension of the standard Sugeno style inference to handle reasoning with multiple instances. In addition to the MI-FIS inference styles, one of the main contributions of this work is an adaptive neuro-fuzzy architecture designed to handle bags of instances as input and capable of learning from ambiguously labeled data. The proposed architecture, called Multiple Instance-ANFIS (MI-ANFIS), extends the standard Adaptive Neuro Fuzzy Inference System (ANFIS). We also propose different methods to identify and learn fuzzy if-then rules in the context of MIL. In particular, a novel learning algorithm for MI-ANFIS is derived. The learning is achieved by using the backpropagation algorithm to identify the premise parameters and consequent parameters of the network. The proposed framework is tested and validated using synthetic and benchmark datasets suitable for MIL problems. Additionally, we apply the proposed Multiple Instance Inference to the problem of region-based image categorization as well as to fuse the output of multiple discrimination algorithms for the purpose of landmine detection using Ground Penetrating Radar.

Share

COinS