Connections with curriculum

Questions I answer:
  • What connections do I see between my work at the internship and the material I learned in the core course or the electives for the Digital Humanities minor?
  • How do research and theories match with practice?
  • What theories can be derived from practice?

I am finally working on an actual product! It is a brand new one that I came up with. I saw the company heading towards the music space and thought about a product that would be an innovative addition to this area and be based on emerging technologies. I decided to focus on developing a social AI-generated music app. After a lot of research, I managed to find and understand some open-source AI models that have been specifically trained to produce music based on certain inputs. My idea is to build on top of these models and allow the user to input even more settings (such as genre, tempo, pitch, etc.). The AI in the app will take these settings (and even other music that the user has uploaded) and create a new piece of unique music.

As of now, this is just an initial idea and I have already started conducting user research and user interviews with people from various backgrounds. I interviewed people in the music space (some artists and producers) along with people with no musical background, just to have a diverse data set. I will have to do a lot more research and ideation in order to understand whether this a product that consumers will even want. I plan to conduct more marketing research and further analyse current technological trends in the music industry. This whole process and the work I am currently doing is very similar to what I have done in other Digital Humanities courses (specifically 101 and 199). There is a lot of research, data analysis, specification write-ups and product development done in my role, which is very similar to the group project I did in DH 101 (which emphasized teamwork and cross-functional collaboration).

I am using tools that we used in DH 101 like Tableau and Voyant Tools to analyse text and data. In fact, I have started creating Tableau visualizations based on the user surveys I have already sent out, just to make it easy for the team and I to see how users are thinking about my potential app and what other problems or pain points they currently have with other social and music apps, and in the music industry in general. I am obtaining a lot of actionable insights from this data analysis and visualization that will help me set KPIs and OKRs for this product, further building on a new data-driven product strategy. Moreover, I will have to create wireframes and later on prototypes for this product, which I also had experience with during DH 101. I definitely learnt a lot of practical skills in my Digital Humanities coursework that I am applying now in my internship!

Furthermore, I learnt about different types of machine learning algorithms during DH 199 as well as employing R and Python for data analysis. I am definitely grateful for that because I am able to apply that knowledge in industry. Understanding the AI music models and researching ways to improve them has definitely been challenging, since the idea that I am proposing requires a lot of work and innovation. Research and theories that I have done and proposed have definitely not yet matched with practice. From the little work I have done in trying to train my own music model in Python using various machine learning algorithms, this is definitely a very difficult problem to deal with. Musical data is much different to text or numbers, and requires so much more processing, even to just input it into the model without any modifications. From my experimentation so far, I have developed a few initial theories regarding how to do this, but I need to do a lot more research and reading, so that I can fully understand the nature of musical data and AI music models.

Leave a Reply

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

You may use these HTML tags and attributes:

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>