MAJOR: Modeling Musical Improvisation to Support Creativity in Education and Performance

“该奖项是根据2009年美国复苏和再投资法案（公法111-5）资助的。“音乐创作是普遍的，至少数万年来一直是人类创造力的渠道，音乐在人类的社会联系，情感交流和娱乐中发挥着重要作用。 由于信息技术和娱乐在发达国家的总经济产出中所占的份额越来越大，创造力是现代经济的核心。 然而，人们对音乐领域的创造力却知之甚少。 在这个项目中，PI和他的团队将寻求理解，建模和支持音乐背景下的即兴创作或实时协作创造力。 世界上大多数音乐传统都将即兴创作作为创造力的一种方法，因此对高度进化的音乐系统中的即兴创作进行分析和建模应该为创造性活动提供必要的见解。 本研究将考虑音乐传统的代表性子集，以保持结果尽可能广泛适用：爵士乐，印度古典音乐和前卫艺术音乐。 该研究将采用跨学科的方法，涉及民族志，音乐理论，统计建模，机器学习，信号处理和乐器设计以及认知研究。 目标是开发即兴创作的计算模型，并利用它们开发支持音乐和教育创造力的新技术。 为此，PI和他的团队将研究音乐即兴创作的机器建模，作为理解实时创造力的一种手段。 他们将对来自许多文化的即兴传统进行分析，并试图使用一个共同的概率框架将它们统一起来。 他们还将在现实的表演环境中对正式模型进行系统的评估，并使用即兴音乐家的大脑成像来深入了解高度创造性的心理活动。 民族志将被应用于描述即兴传统的多样性，而音乐理论将有助于定义基本概念，并为正式模型提供工作假设和框架。 音乐概念之间的关系将使用概率生成模型来表示，该模型旨在表示音乐参数之间以及表演者之间的复杂条件依赖关系;这种抽象将有助于将不同的表面传统统一在一个共同的框架内，揭示系统中实时创造力的基本模式。 通过建模，这项研究将探讨即兴音乐家如何学习这些系统，并在表演中参考它们。 认知研究将阐明即兴创作过程中是否采用了一组独特的认知过程;了解神经网络的活跃程度可能是工程系统在其他环境中培养创造力的第一步。 更广泛的影响：研究创造力得到发展、完善和系统化的艺术，将为开发在工程和信息技术等其他领域培养创造力的系统提供见解。 从这项研究建立的即兴系统将被用来创建合奏，通过编程，计算建模和创造性的问题解决可能会教。 支持即兴创作的技术的发展也将有助于教授创造力以及机械的表演技能，并鼓励新的编程范式。 对于艺术家来说，项目成果将使新的即兴系统能够使用可重新配置的构建块来构建，从而促进即兴研究和复杂系统的开发。 与世界知名音乐家的计划合作将使广大公众了解这项研究，并将有助于招募新的工程学生，将公众的注意力转向创造力的研究，并希望创造令人满意的艺术作品。

"This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5)."Music-making is universal and has been a conduit for human creativity for at least tens of thousands of years, and music plays an essential role in human social bonding, emotional communication, and entertainment. Because information technology and entertainment constitute an increasing share of the total economic output in developed countries, creativity lies at the heart of the modern economy. Yet the creativity that underlies musical domains is poorly understood. In this project the PI and his team will seek to understand, model, and support improvisation, or real-time collaborative creativity, in the context of music. Most musical traditions in the world use improvisation as a method of creativity, so analysis and modeling of improvisation in highly evolved musical systems should provide essential insights into creative activity. This study will consider a representative subset of musical traditions, in order to keep the results as broadly applicable as possible: jazz, Indian classical music, and avant-garde art music. The research will employ an interdisciplinary approach involving ethnography, music theory, statistical modeling, machine learning, signal processing and instrument design, and cognitive studies. The objectives are to develop computational models of improvisation and to use them to develop new technologies that support creativity in music and education. To these ends, the PI and his team will investigate machine modeling of musical improvisation as a means to understand real-time creativity. They will perform analysis of improvisational traditions from many cultures and attempt to unify them using a common probabilistic framework. They will also conduct systematic evaluation of formal models in realistic performance contexts, and use brain imaging of improvising musicians to gain insight into highly creative mental activity. Ethnography will be applied to characterize the diversity of improvisational traditions, while music theory will help define basic concepts and provide working hypotheses and frameworks for formal models. Relationships between musical concepts will be represented using a probabilistic, generative model that seeks to represent the complex conditional dependencies among musical parameters as well as among performers; this abstraction will help unify different surface traditions within a common framework, revealing essential patterns in systems for real-time creativity. Through modeling, the study will explore how improvising musicians learn these systems and reference them in performance. Cognitive studies will elucidate whether a distinct set of cognitive processes are employed during improvisation; understanding what neural networks are active may be a first step in engineering systems that foster creativity in other contexts. Taken together, these studies will improve our understanding of creative processes.Broader Impacts: Studying the arts, where creativity has been developed, refined, and systematized, will offer insight into developing systems that foster creativity in other areas such as engineering and information technology. Improvisation systems built from this research will be used to create ensembles through which programming, computational modeling, and creative problem solving may be taught. The development of technology to support improvisation will also help teach creativity alongside rote performance skills, and encourage new programming paradigms. For artists, project outcomes will enable new improvisation systems to be built using reconfigurable building blocks, thereby catalyzing research in improvisation and the development of sophisticated systems. Planned collaborations with world-renowned musicians will expose a broad public to this research, and will help recruit new students to engineering, turn public attention to the study of creativity, and hopefully also create satisfying works of art.