Holly Grimm
has successfully completed an online offering of
Creative Applications of Deep Learning with TensorFlow III
Taught by
Parag Mital
Course Run Dates: November 1, 2017 — April 30, 2021
Issued: January 4, 2018
Learning Outcomes
Deep Natural Language Processing
- Ability to preprocess words and sentences using NLTK
- Ability to model words using GloVe or Word2Vec's SkipGram and CBOW models
- Ability to model sentences using bucketed or dynamic sequence-to-sequence models
- Ability to model attention in sentences
- Ability to build chat bots, conversational AI, translate language, or encode meaning in rich natural language corpora
Deep Autoregressive Image and Audio Modeling
- Ability to use PixelCNN to model image distributions
- Ability to use WaveNet to model sound distributions
- Ability to use infer with fast generation of image and audio using queues
- Ability to use NSynth to autoencode audio with WaveNet decoding
Deep Generative Music Modeling
- Ability to preprocess MIDI for Google's Magenta library
- Ability to use Google's Magenta library to build generative MIDI
- Ability to model monophonic, polyphonic, improvisational, and drum MIDI