I conduct research on visualization for and with machine learning, mainly to support non experts in using and developing new machine learning models. Currently, I am working on the topic of music composition and how AI and visualization can enhance the human instead of replacing it. One of the answers is: Using a human centered, AI-assisted approach with visualization as efficient communication between human and AI. At the start I focus only on monophonic melody composition. Later I want to expand my research to other steps in the composition process, machine learning pipeline, and human-AI interaction to foster the human creativity, music writing, and bring AI to non expert people.