After spending my sabbatical year in Singapore, I am back in Ithaca. I love this place but could do without the mid-April snow flurries. The good news is that I still have a few months left until I have to start teaching again in August. The bad news is that I don’t have a clear vision for how to best spend this precious time.
During my sabbatical, I had a lot of time to reflect on the successes and failings of my eight-year-long MegsRadio.fm project. The summary is that it is very difficult to create a professional-grade music recommendation service with a rotating cast of (even the most talented) undergraduate students. So I decided to back off of this project and look for something new to explore.
Here is my short-to-medium-term list of potential projects:
- Interview Local Music Experts to figure out what if any role music recommendation technology can be used to grow a local music scene.
- Create a Porchfest Artist Recommendation Mini App to help people discover talent local artists based on a limited amount of available data.
- Compete in the RecSys Grant Challenge with Spotify Playlist data to see how our Latent Markov Embedding technique stacks up
- Learn more about Counterfactual Learning to make better use of logged data from recommender systems.
- Continue to work on our SLIONS Karaoke app which uses singing to help people learn a foreign language.
So the question is how do I decide to allocate my time? I could simply pick the projects that I am most passionate about (1,2), the projects at best align with my expertise (2,3), the ones that stretch me the most (3, 4), the project which is that is closest to completion (5), or the ones that could have the largest potential impact (hard to tell).
For now, I am going to do a bit of all 5 projects and see what sticks. I also have one or two potential consulting gigs that I could pursue but I have never been one to follow the money.