The Machine Learning Algorithm as Creative Musical Tool

Abstract

Machine learning is the capacity of a computational system to learn structures from datasets in order to make predictions on newly seen data. Such an approach offers a significant advantage in music scenarios in which musicians can teach the system to learn an idiosyncratic style, or can break the rules to explore the system's capacity in unexpected ways. In this chapter we draw on music, machine learning, and humancomputer interaction to elucidate an understanding of machine learning algorithms as creative tools for music and the sonic arts. We motivate a new understanding of learning algorithms as human-computer interfaces. We show that, like other interfaces, learning algorithms can be characterised by the ways their affordances intersect with goals of human users. We also argue that the nature of interaction between users and algorithms impacts the usability and usefulness of those algorithms in profound ways. This human-centred view of machine learning motivates our concluding discussion of what it means to employ machine learning as a creative tool.

Embracing First-Person Perspectives in Soma-Based Design

“…Indeed, as designers we can provide the system with examples of behavior or idiosyncratic goals and the system could learn from these examples. For example, in the design of music interaction, a performer-designer can demonstrate to the system personal movements and gestures, and the machine will learn this performer-specific vocabulary [100,101]. This is a powerful tool that can be used by non-technical designers to explore and create designs.…”

Mapping by Observation: Building a User-Tailored Conducting System From Spontaneous Movements

“…Similar applications can be defined in a context were learning does not take place by automatically analyzing spontaneous movements but allowing users to consciously create their own personal mappings. This would be closer to Mapping by Demonstration as introduced by Françoise (2015), and can benefit from Interactive Machine Learning techniques for DMI building (Fiebrink and Caramiaux, 2016). We have in fact already considered this approach in the context of conducting to allow users to define their own space for controlling articulation (Sarasúa et al, 2016).…”

From artificial neural networks to deep learning for music generation: history, concepts and trends

“…In [18], Graves analyzes the application of recurrent neural networks architectures to generate sequences (text and music). In [12], Fiebrink and Caramiaux address the issue of using machine learning to generate creative music. In [51], Pons presents a short historical analysis of the use of neural networks for various types of music applications (that we expand in depth).…”