Building AI Music Composition Products

 Today's talk is the Yin to the Yang associated with yesterday's AI Death Metal talk by CJ Carr of the dadabots group ( or maybe i have that yin-yang labeling backwards).  They are seemingly two very different and contrary viewpoints on how to approach deep learning and apply it to automatic music composition.  But they are interconnected and perhaps even complimentary in our ever more interconnected world.  And it's wonderful they were both presented at the same technical conference on AI and Musical Creativity.


Ed Newton-Rex is a product director at TikTok Europe.  He is working there now because TicTok acquired ByteDance, which acquired Jukedeck (an award winning AI music composition company he founded).


This talk has a very heavy 'business' zeitgeist to it.  We could tag it with terms like serious, market analysis, profit-loss analysis, practical, etc. So it sits at on end of latent space associated with automated music composition and the databots presentation lies in some opposite corner of the latent pace.  

And we'd all like to twist a few knobs and explore the uncharted space in the latent space between these to viewpoints.  Preferably using an automatic AI based system that will do all of the work for us, perhaps even generating the Github posts with all of the code.  

A GAN algorithm to automatically generate different automatic music composition systems for us.


All kidding aside, it's a very interesting talk.  He breaks down the AI automatic music composition marketplace down for us. Various user scenarios, different approaches to monetization, different degrees of user adjustability (or lack of such).

He makes a special point of acknowledging the work of David Cope, who was an early pioneer it he area of Ai based music compositional systems.


The talk also makes a big point of emphasizing choosing one thing to focus on and then trying to do that one thing really well. Hope it was the right one thing.  Doing a few things pretty well might better equip you to adapt to whatever your final potential marketplace and/or product really are, since oftentimes they change once you start getting into really developing a products ad getting feedback from users.

He also points out how in the grand scale of 'where the money is' this is a small potential marketplace (a small sandbox to be playing in).  At least in terms of how we currently think about it.

You may or may not agree with the end results of how he constructed his random forest decision tree for taxonifying potential uses and users of automated compositional systems.  That 'random forest decision tree reference is a nerdy joke, he's not literally using random forest algorithms.  But he is cleaving the world apart in a binary classifier tree like way.


Observation about some people's obsession on potential market size for idea evaluation

Speaking from personal experience, you can build very successful ground breaking software companies in smaller (sometimes very small) potential marketplaces.  Not everyone needs to be the next Google, or the next Tesla.  

It's perfectly fine to create small boutique software (or hardware) companies that focus on doing great work for smaller marketplaces. 

Oftentimes the customers associated with these smaller marketplaces will be super devoted to their niche interest or niche professions.  If you make a good product for these niche users and niche customers, they will love you for it.  Both in terms of buying your software, and then repeat buying it again later via upgrade purchases.  But also by being incredibly thankful that you are devoting energy to listening to them, understanding their unique niche needs, and then trying to address those needs in software you are writing for them. 

Comments

Post a Comment

Popular posts from this blog

CycleGAN: a GAN architecture for learning unpaired image to image transformations

Image
 Like the Doublemint Twins touting the joys of Doublemint gum, 2 GANs are surely better than 1 GAN.  Especially if we package them together inside of one meta GAN module.  And this is exactly what the CycleGAN architecture does. Have you ever harbored dark secrets of turning a horse into a zebra?  The CycleGAN was developed to do just that. Learn how to turn a horse into a zebra.  And more. Now right away you can notice a difference between the image to image transformation GAN architectures we've been discussing over the last few posts.  Those last few posts described systems that learn from a database of matched input-output image pairs. And if your goal is to turn an edge representation into a nicely filled in continuous tone image, it's easy to build your database of matched input-output image pairs that your GAN system can then learn off of.  Take a continuous tone image (which will be the output of the database pair entry), then run it through an edge detector algorithm. 
Read more

Pix2Pix: a GAN architecture for image to image transformation

Image
 I thought following up yesterday's TraVelGAN post with a Pix2Pix GAN post would be useful to compare what is going on in the 2 architectures.  Two different approaches to the same problem. I stole this Pix2Pix Overview slide below from an excellent deeplearning.ai GAN course (note that they borrowed it from the original paper) because it gives you a good feel for what is going on inside of the Pix2Pix architecture.    Note how the Generator part is very much like an auto-encoder architecture, but rebuilt using the U-Net architecture features (based on skip-connections) that fastai has been discussing in their courses for several years before it became more widely known to the deep learning community at large  (and which originally came from an obscure medical image segmentation paper) . So the Generator in this Pix2Pix GAN is really pretty sophisticated, consisting of a whole image to image auto-encoder network with U-Net skip connections to generate better image quality at highe
Read more

Smart Fabrics

Image
"What makes smart fabrics revolutionary is that they have the ability to do many things that traditional fabrics cannot, including communicate, transform, conduct energy and even grow."  This is a quote from Rebeccah Pailes-Friedman, who is a professor at Pratt with a research focus on developing wearable technology and smart textiles. Enhancing textiles can be viewed through different lens.  Enhancements can be aesthetic, or performance boosting, or durability boosting. 
Read more