For the machine learning course, students were asked to explore ‘machines’ that could help users deal with their current situation (COVID-19 pandemic) of social distancing, remote communication and longer time spent at home. With this in mind, the team focused on the question: ‘how can machine ‘friends’ potentially help filling the void?’

Memorem is an artificial friend that ‘lives’ in the user’s computer and brings back memories by recognizing the user’s voice tone and ‘trigger’ words to then subtly display a positive ‘memory’ (image, video, GIF) on the desktop. The project’s intent is to cheer up the user by making them appreciate past memories and the people in them. The team focused on making an unobtrusive ‘machine’ that can make a little positive impact in a person’s life by being present in their most isolated and vulnerable moments.

The machine learning process behind the software uses a voice recognition module in which different ‘trigger’ words are stored. Once it recognizes the word it will select from a library of images, select the one that fits best what the user has said, and then display it as a GIF.

The team was inspired by Mark Weiser’s quote:
“Ultimately, computers would ‘vanish into the background,’ weaving themselves into the fabric of everyday life until they are indistinguishable from it.”