CSCE 496/896 Topic Survey/Implementation Ideas


For a course project, more breadth is expected from a report (literature reivew) than in an implementation. In both cases, a detailed description of different approaches is expected, as well as a careful comparison and a critical assessment.
o Identifying genetic structures that possess certain proteins (implementation; has the possibility of becoming a research project, if desired)

o Principal Component Analysis (PCA) for feature selection (report or implementation)

o Learning in fuzzy systems, e.g. learning a neural network-based fuzzy controller for a network, a robot, or some other engineering system (report or implementation)

o Learning in network control, especially call admission control (CAC) in an ATM network (report or implementation)

o Automated computer system intrusion detection, i.e. automatically identifing when a person has cracked into a computer system (report or implementation; has the possibility of becoming a research project, if desired)

o Learning to classify web pages or usenet news (report or implementation)

o Learning to make content-based queries in a visual database (report)

o Learning in multi-agent systems (report)

o A study of multiple-instance learning, where each labeled example consists of several instances from the instance space, and we don't know which instance is responsible for the label (report, maybe implementation)

o Using reinforcement learning for optimization problems such as TSP (report, maybe implementation)

o A study of new results in boosting and bagging (report, maybe implementation)

o Implementing a version of exponentiated gradient and comparing it to the gradient descent algorithm (implementation)

o Multi-threshold perceptron (report or implementation)

o Using exponentiated gradient in reinforcement learning (report or implementation)

o Using expert-based algorithms (e.g. weighted majority) for pruning decision trees (report or implementation)

o Using expert-based algorithms (e.g. weighted majority) for predicting disk accesses (report or implementation)

o Using expert-based algorithms (e.g. weighted majority) for predicting the stock market or for rating movies (implementation)

o Using statistical hypothesis testing methods for comparing classifiers (report or implementation)

o Dealing with unlabeled data, e.g. clustering, the expectation maximization algorithm, or co-training (report or implementation)

o Learning to learn (report or implementation)

o Learning hidden Markov models [HMMs] (report or implementation)

o Genetic algorithms running in constrained domains (report or implementation)

o Hardware-based genetic algorithms (report)
[an error occurred while processing this directive]
Back

Return to the CSCE 496/896 Home Page