Back to Projects
K
Group Project

KIUA1011: Intro to AI — Final Project

Google Teachable MachineSupervised LearningImage ClassificationDataset CurationAI Product DesignResearch Documentation

Overview

Group project (3 members) for the Intro to AI course at INN. This four-part assignment covered the full cycle of an AI project — from concept design to hands-on model training to ethical reflection. Exercise 1 involved designing a concept for an AI-driven orientation system for the University of Inland Norway, identifying five real student pain points and proposing specific chatbot and recommender solutions, backed by a structured research table with cited sources. Exercise 2 built on this by training an image classification model in Google Teachable Machine using 210 manually curated images across three classes: Cars, Cats, and Dogs. Every image was sourced from Wikimedia Commons under open licenses and documented in a full dataset catalogue with APA 7th citations. The model was tested with two unseen images — a parrot and a motorcycle — to evaluate real-world performance. Exercises 3 and 4 covered ethical and philosophical reflection on AI in education, academic integrity, and Norway's AI regulatory landscape.

My Contributions

  • Designed 5 AI solutions for real student orientation problems at INN, with tool types and cited sources
  • Built a structured Excel research table (Table 1) documenting each activity, problem, and proposed solution
  • Manually curated and licensed 210 images across 3 classes from Wikimedia Commons
  • Documented every image in a full dataset catalogue with APA 7th citations (Table 2)
  • Configured and trained the image model (50 epochs, batch size 16, learning rate 0.001)
  • Tested model with unseen inputs — parrot misclassified as Dog/Cat (54%/44%), motorcycle correctly identified as Car (100%)
  • Wrote ethical reflection on AI use in education, academic integrity, and Norway's AI regulatory landscape

Tech Breakdown

Google Teachable Machine, supervised learning, transfer learning (MobileNet), manual dataset curation, open licensing compliance (CC BY-SA, GFDL, Public Domain), Excel dataset documentation, APA 7th referencing, AI product design thinking

Request Report