AI is heavily intertwined with our day to day lives, for better or for worse. As technology continues to grow and improve, AI’s impact on most industries is, or soon will be monumental.

One of the best known AI examples is IBM’s Watson, a question-answering computer system capable of answering questions posed in natural language. In 2011, Watson competed on “Jeopardy!” against legendary champions, winning the first place prize of $1 million.

Another crucial one is the use of AI in the medical field. We might not be aware of it, but AI technologies are currently used for many aspects of our medical treatment. For example, AI systems can evaluate patient health issues, conducting a continuous, in-depth analysis of complex personal medical data. Machine learning algorithms can also analyze the relationship between a particular form of treatment and the symptoms found, in order for them to understand how to improve the treatment.

So what does technology, and particularly AI have to do with the Music Industry?

Let’s take a step back. The delicate relationship between computers in general and the Music industry started probably around 1951, when famous computer scientist Alan Turing recorded the first-ever computer-generated music (prior to any AI capabilities).

As cool as that may be, the relationship between new technological capabilities and the music industry could really be best defined as “it’s complicated”.

On several occasions, the newest technological inventions came close to killing the Music Industry.  For example, when Radio first came out, or when Napster was introduced to the world (until the recent rise of streaming services) – the Music Industry suffered devastating years. However, despite the difficulties and adjustments incurred by these changes, Music is stronger than anything else, and the Music Industry found a way to recover and outgrow what it previously was

Music Publisher? Check out MyPart for Music Publishers

Despite the differences (and inherent conflicts) between the two, music and tech are best known for their cooperation in three main fields. The first one is providing accessibility to music for the general public in easy and intuitive ways (with the likes of Spotify, Pandora and Youtube).

The second and most talked about field in recent years is the recommendation and suggestions systems for music listeners. These systems help users get acquainted with new and exciting music, based on their own personal taste.

The third one is using innovative technology to help artists and producers create and implement their musical visions more easily, with ground breaking tools (such as Cubase, Pro Tools, Ableton Live, etc).

Recommendation systems were in many ways the first aspect of music and tech that involved the use of AI capabilities. These systems collect enormous amounts of information about songs and listening habits, suggesting similar and relevant songs to follow up with.

However, despite recent progress, AI hasn’t penetrated intro all aspects of the Music Industry. At MyPart for example, we’ve developed an AI based platform for the sole purpose of song search within the music industry. MyPart takes a very different approach when looking at the problem of song placements and song relevance.
We’re able to significantly improve the likelihood of matching any one song with a performing artist or music supervisor on the other end.

We do this by looking at the challenge as a Machine Learning Classification Problem. AI could be trained to solve such problems, when given enough examples to learn from. Our ML model analyzes hundreds of harmonic, melodic, lyrical and structural features, and prioritizes the results for music executives according to current project needs and personal preferences, based on a benchmark of musical references that they feed the system with. This way, we can offer both music publishers and performing artists much better odds at finding the song they’re looking for.

Under the Hood

Every benchmark defined by a music executive (i.e. record label or music publishing representative) defines a scope for the AI to sink its teeth into, in order to understand what are the deep common denominators lyrically and musically. Trained on this benchmark, MyPart’s AI can then analyze any given song, and sort according to ‘relevance’ predictions to every defined benchmark. Results are presented to the music executive prioritized, on a dedicated MyPart mobile app.

MyPart supports two distinct flows:

  1. Song Discovery within massive Music Publisher catalogues
  2. Song Discovery for the mass audience: the huge pool of undiscovered talent out there. Anyone can now submit a song/lyrics to a target of their choice on MyPart’s platform. MyPart is officially collecting songs this way for leading performing artists including Bruno Mars, Kelly Clarkson, Aerosmith, and many more.

At MyPart, we strongly believe that the connection between technology and specifically AI capabilities will benefit the music industry in many meaningful ways. Our humble purpose is to make the jobs of music industry executives, publishers and performing artists significantly more effective, by helping them focus on relevant content from the ever growing avalanche of potential songs coming in, whether from music publishing catalogues or aspiring unsigned artists from around the globe.

By: Sivan Kollnescher & Zack Martin

Leave a Reply

Your email address will not be published. Required fields are marked *