While AI technology is designing a future where we have better communication with the robots, scientists are vividly faced with the problem of translating intrinsic human intuitions and complex attributes of human into computer codes and figures. The human is a subject of varied complex emotions; different family backgrounds, different psychology and social creature. These and more make life very interesting but pose extreme difficulties in developing intelligent technology models that would interact swiftly with the humans. An award-winning designer, artist, and researcher, Raphael Arar thinks we can teach a computer to make sense with arts and nostalgia scores.
The Challenge at a Glance
Think about how you felt hearing your favorite song for the first time. Can you describe in full context how it made you feel? Did you get fired up or get Goosebumps? That’s obviously hard to describe. “Part of us feels so simple, but under the surface, there’s really a ton of complexity,” Arar said at TED talk.“And translating that complexity into machines is what makes the modern day’s moon shots”. There are lots of human emotions; qualities which our fingers obviously cannot access.
Arar is convinced that translating these emotions that we cannot really describe to each other is not what we can answer using just 0s and 1s alone. He is taking arts as a way of designing better experiences that can be used to develop AI systems which would have better communications with the humans. And his progress seems so, as he said it’s been serving as a catalyst to beef up the most human ways that computer can relate to us.
“I view art as the gateway, to help us bridge this gap between humans and machines,” Arar said.“Through arts, we are tackling some of the hardest questions like: what does it really mean to feel? Or how do we engage and know how to be pleasant with each other? And how does intuition affect the way that we interact?”
Currently, AI technology has been developed to make sense of some basic human emotions, like fear, sadness and joy. This was achieved by simply converting those characteristics into maths. But the challenge is on the complex emotions we have a hard time describing to each other, like nostalgia.
Arar worked with some data scientist to explore the complex human emotions and to establish a model that can be used to convert them into something mathematical. He created a piece of art; an experience which asked people to share a memory and a model he referred to as “nostalgia score.” The stories shared were analyzed using the computer for a simpler emotion, considering words that are associated with nostalgia in the story and the result provides a nostalgia score. Arar disclosed that the score is used to build a light-base nostalgic sculpture that serves as the physical embodiment of the person’s contribution. “The higher the nostalgia score, the rosier the hue,” and the score “describes how the experience made you feel.”
Breaking down the steps involved in having a conversation with some may help us to understand how much we need to teach AI systems before they can interact with people well. Scenarios like knowing when to change the topic or even choosing the topics to discuss are like a second nature we have not accounted for in AI technology. We should rather teach the computer how to make sense of the emotions we have hard times explaining to each other.