Abstract
Evolution is often summed up as “survival of the fittest”. Evolutionary computation borrows this logic to solve hard optimization problems, from designing robots to building interpretable models. But evolution in nature does more than optimize. It keeps many diverse forms of life alive, and some of its biggest leaps came from cooperation rather than competition.
In this talk, I will give a non-technical tour of evolutionary computation, focusing on diversity, and show how these ideas turn evolution from a problem solver into a creative process. I will then present an early-stage project at RITMO on interactive rhythm exploration. I welcome feedback and ideas from the RITMO community.
Bio
Berfin Sakallioglu is a PhD candidate in Applied Data Science and Artificial Intelligence at the University of Trieste, Italy, currently visiting RITMO, hosted by Kyrre Glette. Her research is in evolutionary computation, focusing on interpretability, diversity and learning in evolutionary systems, including hybrids with large language models. At RITMO, she explores how evolution can support musical creativity. She completed her BSc in electronics engineering in Ankara, Turkey, and her MSc in data science in Lisbon, Portugal, where she also taught at the university. Outside of research, she performs as a singer with her acoustic guitar, and has a deep interest in cultures, languages, biology, psychology and neuroscience. To learn more about Berfin's work visit berfins.github.io.