Implemented and trained deep attractor network from ‘Deep attractor network for single-microphone speaker separation’ in tensorflow ane pytorch for audio source separation.
Tried to correlate sources with listeners EEG signal with Max Jakobs.
Bandgaps
Developed ML model to predict bandgaps of materials given their chemical formula.
Won hackathon (10 teams of Cambridge PhD students).
What’s the title?
Developed and trained bidirectional LSTM model to predict titles of academic papers based on the abstract with Thomas Parton
Presented at conference and won competition with entry.
