Prerequisites

In all cases, we will try to reintroduce and review key concepts, but participants are unlikely to be able to follow the tutorial content in depth at the speed at which it necessarily proceeds unless they already have a reasonable understanding of the below concepts.

• • We assume that people are familiar with basic machine learning, and understand ideas like loss functions, logistic regression, and parameter optimization. One source for this material is the first three weeks of Andrew Ng’s online Coursera Machine Learning class, https://www.coursera.org/learn/machine-learning

  • We generally assume that people have some experience with the basics of neural networks, such as how to train a simple feed-forward network, and of distributed representations, such as word vectors. One source for this material is Michael Nielsen’s online book Neural Networks and Deep Learning, http://neuralnetworksanddeeplearning.com/ . We will explain neural language models and recurrent neural networks in detail.

  • We will assume that people are generally familiar with machine translation and phrase-based statistical MT systems. A good source for this material is: Adam Lopez. 2008. Statistical Machine Translation. In ACM Computing Surveys 40(3),https://alopez.github.io/papers/survey.pdf