Skip to main content Skip to navigation

TMIC: App Inventor Extension for the Deployment of Image Classification Models Exported from Teachable Machine

Project Overview

The document explores the use of generative AI in education through the TMIC extension, which facilitates the deployment of machine learning (ML) models created using Google Teachable Machine within K-12 settings. This user-friendly interface allows students and educators to build and implement image classification models in the App Inventor programming environment, thereby promoting a hands-on approach to learning about ML concepts. By empowering students to create technology rather than merely consume it, the initiative emphasizes the significance of ML education in light of its increasing relevance in everyday life. Through this approach, the document illustrates how generative AI can enhance educational experiences by fostering creativity, critical thinking, and technical skills among young learners, ultimately preparing them for a future where technology plays a central role. The findings suggest that integrating such tools in K-12 curricula can significantly enrich students' understanding of machine learning and its applications, paving the way for a more informed and capable generation.

Key Applications

TMIC (Teachable Machine Image Classifier) Extension

Context: Used for teaching ML in K-12 schools and introductory courses in higher education, also suitable for anyone interested in creating intelligent applications.

Implementation: The extension is integrated into the App Inventor framework, allowing users to deploy trained models as part of mobile applications. It supports TensorFlow.js models exported to Google Cloud.

Outcomes: Facilitates the understanding and application of ML concepts, encourages creativity in app development, and provides a practical way for students to learn about AI technologies.

Challenges: Currently supports only TensorFlow.js models from Google Cloud and is limited to using the smartphone's rear-facing camera for image classification.

Implementation Barriers

Technical Limitations

The TMIC extension currently only supports TensorFlow.js models and has limited functionality for capturing images.

Proposed Solutions: Future work aims to improve the extension to support more model types and functionality.

Project Team

Fabiano Pereira de Oliveira

Researcher

Christiane Gresse von Wangenheim

Researcher

Jean C. R. Hauck

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Fabiano Pereira de Oliveira, Christiane Gresse von Wangenheim, Jean C. R. Hauck

Source Publication: View Original PaperLink opens in a new window

Project Contact: Dr. Jianhua Yang

LLM Model Version: gpt-4o-mini-2024-07-18

Analysis Provider: Openai

Let us know you agree to cookies