Deep Learning Course (Autumn 2020):

second part

Registration will open soon

 

The second part of the course is designed for those who already  have experience with neural networks. In particular, we invite everyone who has passed the first part. The main topics of the course are word and audio processing.

Class format

Classes are held in the format of webinars on the Twitch platform.

Completion and delivery of homework are organized  on the Stepik platform

Program
(September-February 2021)

Below are the topics present in the last run of the course.
The new launch program might change slightly.

  1. Introduction to NLP. Word embeddings

  2. Recurrent neural networks and text classification

  3. Language models

  4. Task  Seq2Seq and machine translation

  5. Modern NLP architectures: BERT, Transformer, ELMO

  6. Dialogue systems

  7. Text summarization

  8. Introduction to Audio Signal Processing

  9. Speech recognition

  10. Speech generation