Congratulations Dr. Jan Melechovsky on obtaining your PhD! I’ve had the pleasure to guide Jan through his PhD journey at the AMAAI lab at Singapore University of Technology and Design. The last five years, Jan has explored a number of fascinating (yet connected ; ) topics, ranging from dysarthric speech analysis to text-to-music.

Some highlights:

Mustango: a controllable text-to-music system - one of the first open systems to directly generate controllable music as audio. More info and listen to samples here.

MidiCaps: first large-scale dataset on text captioned midi files. Download the dataset here.

Dysarthric speech: in a collaboration with the Czech Technical University in Prague, Jan created deep learning models for the assessment of oral diadochokinesis among progressive neurological diseases, dysarthria types and severity.

Accented speech synthesis: disentangling speaker identity from accent is no trivial task. Jan developed three models, including DART.

SonicMaster: first all-in-one music mastering and restoration model. Read the paper and try a demo.

Read Jan’s full thesis here.

Thank you to all Jan’s collaborators: Soujanya Poria, Berrak Sisman, Ambuj Mehrish, Abhinaba Roy, Deepanway Ghosal, Nicolas Guo, Tongyu Lu

Good luck in your future endeavors Jan!