llms-in-music-composition | episodes

listen on castbox.fmlisten on google podcastslisten on player.fmlisten on pocketcastslisten on podcast addictlisten on tuninlisten on Amazon Musiclisten on Stitcher


LLMs in Music Composition

In this episode, we are joined by Carlos Hernández Oliván, a Ph.D. student at the University of Zaragoza. Carlos’s interest focuses on building new models for symbolic music generation.

Carlos began by explaining how machine learning helps music composers improve their composition. He shared the standard data formats in which music is represented for machine learning. He also discussed the machine learning algorithms best for music composition.

Carlos discussed the datasets used for training music composer models. He explained how the model truly learns to generate harmonic notes. He also talked about how these models are evaluated. He also discussed how composers can fine-tune the model to improve the generated music output.

Carlos shared his thoughts on whether these models are genuinely creative. He revealed situations where AI-generated music can pass the Turing test. He also shared some essential considerations when constructing models for music composition.

Carlos’s next research is on symbolic music generation with human emotions. Follow Carlos on Twitter @carlosheroliv.

Carlos Hernández Oliván

Carlos Hernández Oliván is a PhD candidate at the Universidad de Zaragoza. His interests regard Artificial Intelligence and its applications in the fields of music and audio. He studied viola at the Professional Conservatory of Zaragoza, and he completed his Bachelor’s and Master’s degrees in Industrial Engineering at the University of Zaragoza. During his PhD, he interned at Sony R&D in Tokyo, Japan, and he recently received the Vulcanus grant to intern at NTT CSL in Japan. For discussions about music and audio research, please contact me at carlosholivan@gmail.com