For years I’ve wanted to build a super-dense electronic-music compressor: keep only the loops and phase cues that really matter, then re-synthesise the track perfectly. Evenings and weekends, however, were never long enough to design the model, write the maths, and wrangle PyTorch. Recently I opened ChatGPT running the …

Recently, I built a real-time voice Q&A assistant for a working spa business — designed to answer calls, provide instant information, and send helpful follow-ups via SMS. It uses OpenAI’s GPT-4o Mini real-time API for language and voice capabilities, and integrates with Twilio for telephony and messaging. It’s …

I'm excited to share the latest updates to the VSound App, our real-time audio visualization tool designed for musicians, language learners, and audio enthusiasts. 🚀 What's New M1 Mac Compatibility: The app now runs smoothly on Apple's M1 Macs, ensuring better performance and efficiency.​ Performance Improvements: Enhanced responsiveness and reduced latency …

Modern electronic house and techno music is highly repetitive. I am not saying that it is, therefore, boring. I am a big fan of certain styles myself! What I am saying is that patterns are being repeated over and over, and it is the variation of the patterns over time …

I’ve decided to test a small (not very deep) autoencoder on the audio data. It has two convolutional layers, with convolutions of size 4 and strides of 2, with ReLU nonlinearities. This results in 1/4 frame rate of the original audio. The corresponding deconvolutions are then applied. The …

I was asked at the interview what’s the difference between LDA and document clustering. I tried to explain it by explaining the difference between generative models that are assumed for the respective models. However, now I realise it would have been much more effective to give a much simpler …

Released version 0.0.2 of the Java source code. This release allows specifying LDA topic keywords. Please see the software page for downloads.

I present a case study on a corpus of 10,000 news articles. We will investigate the topic structure of the corpus, by gradually “freezing” the topics through specifying their keywords, and seeing what other topics come up. The process shows how you can extract useful topics from your corpus …

The first version of the Java source code for running the heavily optimised LDA Gibbs sampling has been released. Please see the software page for details. Below you can see the 20 topics output produced from 10,000 news articles, after running just 200 iterations. First 100 iterations are burn-in …

When estimating the parameters of the LDA (Latent Dirichlet Allocation) model using Gibbs sampling, if we set the Dirichlet priors to the fixed target values (usually small), then we reduce the mixing of the samples from the target distribution from the beginning, even though we haven’t found a good …