Recently presented our short-paper “CycleMix: Mixing Source Domains for Domain Generalization in Style-Dependent Data” in SETN 2024. Feel free to take a look at the preprint and code.