Synesthetic Algorithmic Composition: A Bio-Inspired Framework for Cross-Modal Creative Generation

Posted: Oct 28, 2025

• Dr. Elara Vance
• Prof. Kenji Tanaka
• Dr. Isabella Rossi

Abstract

The intersection of computational creativity and multi-modal perception represents a frontier in artificial intelligence research with profound implications for both artistic practice and cognitive science. While significant advances have been made in generative systems for individual creative domains, the computational modeling of cross-modal creative expression remains largely unexplored. This paper addresses this gap by introducing a novel framework for synesthetic algorithmic composition that generates intrinsically linked auditory and visual artworks through bio-inspired neural mechanisms. Synesthesia, the neurological phenomenon where stimulation of one sensory pathway leads to automatic experiences in a second pathway, provides a compelling model for computational creativity. Previous approaches to multi-modal generation have typically employed separate systems for different modalities with post-hoc alignment, resulting in outputs that lack the deep semantic integration characteristic of human cross-modal perception. Our work fundamentally rethinks this paradigm by developing computational models that mirror the neural underpinnings of synesthetic experience. We propose three research questions: (1) How can we computationally model the structural mappings between auditory and visual creative domains? (2) What architectural principles enable the generation of coherent multi-modal artistic outputs? (3) To what extent do such computational models capture meaningful aspects of human cross-modal perception? Our contributions include: the Cross-Modal Resonance Network architecture, a novel dataset of aligned musical and visual artworks, and empirical validation through both computational metrics and human evaluation.