A3: Class Presentation – Group Project

Course CMPT 310 D100
Semester Summer 2012
Title Artificial Intelligence Survey
Campus Burnaby Campus
Instructor(s) Steven Pearce <stevenp@sfu.ca>
TA(s) Sitanshu Gakkhar <sgakkhar@sfu.ca>
Criteria criteria.pdf
Students(s) David Carlson, Davie Ching,
Mark Lauman, Christopher Kwong

Topic

Advanced Learning Agent Analysis

Abstract

We will be presenting an advanced topic on a learning agent implementation; by specifically focusing on the agent’s implementation by way of neural networks and genetic algorithms. Our goal is to provide the class with a tangible example of how neural networks can be implemented and how to artificially simulate learning with a defined environment. The presentation will consist of a demonstration of the problem domain, analysis of the neural network and genetic algorithm implementations, and will culminate in critical analysis of the base example. Within the critical analysis, we hope to simulate changes to the problem domain in the form of manipulating the neural network architecture (ie. increase/decrease number of neurons within the network) and externally controlling the environment (ie. limit goal availability).

Our base case example upon which our analysis and implementation will be upon can be found at: http://site.nixuz.com/evolving-fish-intelligence

Project Submission

Resource Contents Description Filetype
Presentation Presentation Presentation used in class. Powerpoint (*.pptx)
Analysis Code Analysis of the source code. Acrobat (*.pdf)
Graphs Analysis of the fish over time. Acrobat (*.pdf)
Program 1x6 (Original) 1 row, 6 columns network. Version used in base case. Webpage (*.html)
2x6 2 rows, 6 columns network Webpage (*.html)
3x2 3 rows, 2 columns network Webpage (*.html)
6x1 6 rows, 1 column network Webpage (*.html)
6x6 6 rows, 6 columns network Webpage (*.html)
11x24 11 rows, 24 columns network. Processor intensive! Webpage (*.html)