Feasibility Study with Edinburgh Innovations: Scaling Solutions for Neural Network Training with Artist IP Protection

人工智能（AI）正在从根本上改变我们定义、生产、销售、体验、保护和分享音乐的方式。DataMind Audio是一家由音乐家为音乐家创建的新公司，得到了AHRC资助的创意信息学项目的宝贵支持。（爱丁堡大学）。在这项与爱丁堡创新有限公司合作进行的可行性研究中，由里德音乐学院（爱丁堡大学）院长马丁·帕克博士领导，DataMind Audio将与电子音乐家合作，评估用于训练我们的神经网络或“艺术家大脑”的缩放解决方案。基于插件或电子仪器允许用户将任何输入声音输入到根据艺术家的创造性输出训练的“艺术家大脑”中。我们的“艺术家大脑”从道德上获取原创艺术家风格的本质和模式。目前，我们的算法是根据20位世界级艺术家的创意作品进行训练的，其中包括理查德·迪瓦恩、马克斯·库珀和比尔先生。这些算法为声音设计师解锁了新的音景，以全新的方式探索和分享。我们在声音设计和人工智能领域的著名艺术家和先驱名单证明了我们奋进的质量，原创性和诚信。我们最近为期一个月的人工智能研讨会系列得到了Tipper和Slander等电子音乐制作明星的支持，以及Ableton用户研究主管Till Richter等经验丰富的领域领导者的支持。我们的创始人团队已经建立了一个突破性的产品系统，保护原创艺术家的知识产权，并为他们的“艺术家大脑”支付版税，这些神经网络是根据音乐输出训练的。我们使用人工智能通过一个道德的英国在线市场来增加而不是减少音乐家的收入来源。作为规模化的重要一步，爱丁堡的帕克博士与DataMind Audio一起进行了一项可行性研究，以便：* 开发一个具有成本效益的标准化工作流程，用于训练我们的神经网络或“艺术家大脑”;* 测试一个以人工智能为重点的工作规范，供4名“艺术家大脑”的人类训练师使用，我们称他们为模型可靠性工程师（MRE）;* 吸引一批下一代音乐家/声音设计师，利用爱丁堡大学相关教学课程的毕业生，培养“艺术家大脑”，并向他们介绍神经网络的进步;* 评估我们从这个项目中获得的发现，以评估复制和扩大规模的最佳方式，同时保留真正的人类策划的音乐质量-我们产品的关键区别。

Artificial Intelligence (AI) is radically altering the way we define, produce, sell, experience, protect, and share music. DataMind Audio is a new company created by musicians for musicians with the invaluable support of the AHRC-funded Creative Informatics project. (University of Edinburgh).During this feasibility study, conducted in partnership with Edinburgh Innovations Ltd, led by Dr. Martin Parker, Head of the Reid School of Music (University of Edinburgh), DataMind Audio will work with electronic musicians to evaluate a scaling solution for training our neural networks or 'Artist Brains'.Our AI-based plugins or electronic instruments allow users to feed any input sound into an 'Artist Brain' trained on the creative output of an artist. Our 'Artist Brains' ethically access the essence and patterns of an original artist's style. We sell these online from our UK-based marketplace.Currently, our algorithms are trained on the creative output of 20 world-class artists including Richard Devine, Max Cooper, and Mr. Bill. These algorithms unlock new soundscapes for sound designers to explore and share in totally new ways. Our list of notable artists and pioneers in the field of sound design and AI speaks to the quality, originality, and integrity of our endeavor. Our recent month-long AI workshop series was supported by electronic music production stars such as Tipper and Slander, as well as seasoned field leaders like the head of User Research for Ableton, Till Richter.Our founder team has already built a breakthrough product system that protects the intellectual property of original artists and pays them royalties for the use of their 'Artist Brains', which are neural networks trained on their musical output. We use AI to increase, rather than curtail, musicians' income streams via an ethical UK-based online marketplace.As a vital step in scaling, Edinburgh's Dr. Parker has structured a feasibility study with DataMind Audio in order to:* Develop a cost-efficient, standardized workflow for training our neural networks, or 'Artist Brains';* Test an AI-focused job-spec for 4 human trainers of the 'Artist Brains', who we call Model Reliability Engineers (MREs);* Engage a group of the next generation of musicians/sound designers, drawing on graduates of the relevant University of Edinburgh teaching programmes, to train the 'Artist Brains', and introduce them to advances in neural networks;* Evaluate our findings from this project to assess the best ways to replicate and scale up, while retaining authentic human-curated musical quality -- the key differentiator of our product.