JohnTan_Project1_Reflection.pdf - Project 1 Reflection 1...

This preview shows page 1 - 4 out of 13 pages.

We have textbook solutions for you!
The document you are viewing contains questions related to this textbook.
Illustrated Microsoft Office 365 & Office 2016 for Medical Professionals
The document you are viewing contains questions related to this textbook.
Chapter PPM2 / Exercise 20
Illustrated Microsoft Office 365 & Office 2016 for Medical Professionals
Beskeen/Cram
Expert Verified
Project 1 Reflection 1 John Tan Professor Ashok Goel CS 7637: Knowledge-Based AI February 10, 2018 Project 1 Reflection Agent Reasoning: Agent’s reasoning approach – Visual or Verbal? The agent I created is mainly visual approach that is similar to the way human’s use logical reasoning to solve a Raven’s Progressive Matrix problem. It’s the same way that a human look at the missing piece based on the pattern of object A to B, horizontally or A to C, vertically. The agent uses the generate and test approach to solve RPM because it uses the transformation between A and B or A and C to generate the answer for D, depending on the coefficient correlation conditions between the images and compare it to 1,2,3,4,5, and 6 individually to see which one closely match to D. The semantic network would be a better approach for verbal but since my agent uses PIL for image processing functionality to open and load images, it only uses generate and test along with affine methods, such as transformation horizontally and vertically and different angle of rotations. The only time it uses the verbal is when the agent obtain the object figures, such as A, B, C, 1, 2, 3, 4, 5, and 6 for 2x2 problem types. Agent’s data representation and problem solving: Agent’s representation and overall problem-solving process. My agent creates objects of visual Raven’s Figures that comprise of its name and list of objects. It represents images efficiently by using the Image class from PIL to load and process images. Then, it measure how similar two images are from A to B and A to C by calculating the root-mean-square difference from the histogram of the
We have textbook solutions for you!
The document you are viewing contains questions related to this textbook.
Illustrated Microsoft Office 365 & Office 2016 for Medical Professionals
The document you are viewing contains questions related to this textbook.
Chapter PPM2 / Exercise 20
Illustrated Microsoft Office 365 & Office 2016 for Medical Professionals
Beskeen/Cram
Expert Verified
Project 1 Reflection 2 images in gray scale, which are a list of pixel counts. Before it calculates the root- mean-square difference to compare to two images of A to B, then C to 1,2,3,4,5, and 6 or A to C, then B to 1,2,3,4,5,6, it does the transformation of rotating the angles horizontally or vertically, or it does the transformation only by its horizontal reflection or vertical reflection. 1. Problem Types – Only getting the 2x2 problem types 2. Visual Raven’s Figure – Get the figures and solution figures via the ProblemSetList text 3. Methods – These methods are used for image processing, converting image to gray scale, transformation based the angle rotated required and as well as the reflection horizontally and vertically, and finally compute the similarity between the images using root mean square. The objects of the visual raven’s figures comes from a concept call frames. The values in the slot in each frame are the attributes. However, since we only using visual approach here, we are only using PIL for image processing and as well as the root mean square to compare the different figures to get the closest similarity.
Project 1 Reflection 3 Agent Design: Overview of the design My agent performs the affine method of transformation, rotation, and flips or reflects horizontally or vertically in 2D along with generate and test by finding the

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture