UWB Logo

Computer Science and Software Engineering Capstone Presentations

Summer Quarter

August 21, 2020

Kayla Sprague

"Building a Mind"

(UWB CSS Faculty Research)

Faculty Advisor: Dr. Yusuf Pisan

Abstract

Building A Mind is a project that originated with the goal of applying a modern approach to solving traditional artificial intelligence (AI) problems. Problems we solve are derived from "Building Problem Solvers," which implements the problems in LISP. LISP is an old programming language that, while commonly used for AI problems, is improved on by current programming languages like Python. To solve the problems, we use current tools to reframe the problems with a reasoning system implemented in Python. A reasoning system is a framework for processing facts about a given problem, applying rules that constrain the solution, and find a solution path based on the initial state and the goal state of a problem. We first used PyKE, a knowledge-based inference engine that integrates with Python. PyKE allowed us to set up the problems with three types of knowledge bases - fact, rule, and question - and provided pattern matching functionality to solve the problem. However, when attempting to solve our first problem, we realized that there are limitations with PyKE. First, debugging PyKE is time-consuming and makes error finding difficult. Secondly, there are syntactical differences with implementation between PyKE and Python. Thirdly, is a significant learning curve in implementing the pattern matching functionality that PyKE provided us. The limitations we experienced with PyKE provoked us to create our own reasoning system. In creating our own reasoning system, we established the same functionality that PyKE provided us, but in our own implemented environment. This collaborative implementation allowed us to increase our knowledge of reasoning systems by building the foundational functionality needed to solve the AI problems. The current state of the system solves few dynamic problems ranging in complexity using a set of reasoners catered to the individual requirements of each problem. Although there are many more problems to be solved and the system is made to be expandable for further implementation of reasoners. This system now serves as an educational opportunity for future students who desire a deep dive into AI by building onto our reasoning system.

Updated August 20, 2020, 17:06