Ruprecht-Karls-Universität Heidelberg
Institut für Computerlinguistik

Bilder vom Neuenheimer Feld, Heidelberg und der Universität Heidelberg

Imitation and Reinforcement Learning for NLP

Module Description

Course Module Abbreviation Credit Points
BA-2010 AS-CL 8 LP
Master SS-CL, SS-TAC 8 LP
Lecturer Artem Sokolov
Module Type Projektseminar
Language Englisch
First Session 13.11.2020
Time and Place Friday, 09:15-10:45, Online
Commitment-Frist tbd.

Prerequisite for Participation

Knowledge of the following will be helpful: foundations of statistical machine Learning, reinforcement learning and neural networks.

Assessment

Implementation project

Content

This module is an software implementation project of imitation learning approaches to sequence-to-sequence learning problems with a focus on neural machine translation. Closely related to structured prediction and reinforcement learning, imitation learning is particularly suited for sequence prediction tasks, where good success metrics or intermediate rewards are hard to define, while in the same time it is easy to provide demonstrations of correct behavior.

Literature

  • Sutton and Barto "Reinforcement Learning" (2018) 2nd edition. MIT Press. http://incompleteideas.net/book/the-book-2nd.html
  • Szepesvari (2010). Algorithms for Reinforcement Learning. Morgan & Claypool. https://sites.ualberta.ca/~szepesva/RLBook.html
  • Osa, Pajarinen, Neumann, Bagnell, Abbeel, Peters (2018), An Algorithmic Perspective on Imitation Learning. Foundations and Trends in Robotics. https://arxiv.org/abs/1811.06711
  • Daumé III, A Course in Machine Learning, Chapter 18 http://ciml.info/
  • Goldberg (2015). A Primer on Neural Network Models for Natural Language Processing. https://arxiv.org/abs/1510.00726
  • Neubig (2017). Neural Machine Translation and Sequence-to-sequence Models: A Tutorial. https://arxiv.org/abs/1703.01619
  • » weitere Kursmaterialien

    zum Seitenanfang