Neural Machine Translation Fall 2017

This is a special topics course entitled “Neural Machine Translation”. This is an advanced graduate seminar course is about recursive neural network models for machine translation (for natural languages such as English, Chinese, Tagalog, etc.). You will require some background in statistical machine translation and language models. This course will not be a tutorial on deep learning in general. The main goal of the course will be to work on writing our own end to end neural machine translation system. The course will cover recent developments in neural machine translation and we hope to discuss and solve issues with the long tail (limited vocabulary size in neural MT) and in the proper modeling of alignment (aka attention).

Instructor

Anoop Sarkar

Office Hours: Wednesday 9:30am-10:30am TASC1 9427

Discussion Board

on Coursys

Time and place

MW 2:30-4:00, AQ 5008

Calendar

Subscribe

Textbook

No required textbook. Online readings provided in Syllabus.

Grading

Activities and grades on Coursys

• In class presentations: 40%
• Discussion board: 20%
• Project proposal: 10% (Nov 19, 2017) use the same submission format as the final project
• Final project: 30%. Due on December 8, 2017

Presentation Sign-up

Sign-up