Smart Coaster: A Friend or Foe?

How can machine learning help us in our day to day lives? What are the ethical implications to consider when designing with machine learning technology?  These were some of the questions the design team set out to explore when working on this project.

In this playful yet thought-provoking exploration, the team centered down to “bad” habits and looked for ways to help nudge people into breaking these habits when willpower falls short.

Noticing that many people around us live a highly caffeinated life, the team created a coffee recognition machine that offers tangible cues within its physical environment, to remind coffee addicts to be mindful of their coffee habits.

The team broke down the workflow into two parts: the input and the output. The input involved training a machine to differentiate between coffee and other beverages using object image detection. Training was done through Wekinator, a supervised machine learning software that allows users to create computational functions by providing training examples of input and output. For the output, the team used a microcontroller to drive two servo motors that brought the coaster to life. Demonstrated in the form of a moving coaster, it is directed away from the drinker when coffee is consumed. It otherwise remains stationary when consuming any other beverage example water.