Reference no: EM133835069 , Length: word count:4000

Concepts in AI

Learning Outcome 1: Appraise a comprehensive/detailed knowledge of Artificial Intelligence

Learning Outcome 2: Design and develop artificially intelligent software artefacts.

Task

Symbolic methods:

1. Develop a semantic network and subsequent frames that describes the gecko family of reptiles. You should offer as much information about one type of gecko, for instance an African Fat Tailed, Crested or Flying Gecko (or any other type of gecko). You should start at the class ‘Reptile' and then define the subclasses that lead to your chosen species of gecko. Each class level should have appropriate properties and you should show at least one instance of your chosen gecko species with its own individual properties. You should aim to present around 5 levels of classes including the ‘Reptile' class. You should state any assumptions made when constructing your semantic network. You will need to research the gecko family and species and sub-species of reptiles, be sure to cite references to any research used

2. Based on your gecko semantic network, describe the appropriate frames, slots, slot-values and default-values. This does not need to be implemented in a programming language, just provide the written definitions of your crow semantic network. Explore and define inheritance in your frames

3. For extra credit, transform your frames into predicate logic expressions

Connectionist methods:

1. Explore the input and output features, attributes and data for the ‘Breast Cancer Care' dataset. You should aim to describe the data in terms of attributes, classes and distribution of classes. You should also explain what the dataset is intended for. Get Fast, Top-Rated Expert Assignment Help!

2. Experiment implementing a model of the Breast Cancer dataset using Weka. You should implementdecision tree, logistic regression and multilayer perceptron models using the dataset. Report the performance for each model, discuss and contrast the results in terms of performance accuracy, the computational complexity of the learning method and the resources (time) used to calculate the outcome.

3. For extra credit, use generative AI (ChatGPT or similar) to generate python code to build a neural network and deep learning neural network classification model for the breast cancerdata set. Your solution should include the classification model, learning curves, confusion matrixes and performance statistics. You should reflect on the performance of the trained model based on the evidence obtained by the statistics, confusion matrixes and learning curves.