• Comparison to Other Work
• Generalizing to more sensor types and machine learning algorithms.
• Enable sharing of data and models (and portability).
• Focusing specifically on audio recognition examples.
• Running on Microcontrollers
• Making Accessible to a More Novice Audience