C S 479
Download as PDF
Introduction to Machine Translation
Course Description
Evolution of machine translation technologies and algorithms, with a foundation in basic algorithms, human-machine interaction, automatic adaptation, statistical and neural models, multilingual models, multimodal models, quality evaluation and estimation, and speech-to-speech translation.
When Taught
Fall
Fixed/Max
3
Fixed
3
Other Prerequisites
Reading ability in a language other than English.
Recommended
Python proficiency, CS 474 or CS 472 or Ling 581 or similar exposure to MT or NLP technologies.
Title
MT Technology Evolution
Learning Outcome
Analyze how the challenges of producing quality translations have been addressed by past and present technologies to identify recurring patterns in linguistic computation. This fulfills the Intellectually Enlarging aim by providing a "broadening" perspective on the historical and technical development of human-machine communication.
Title
Building blocks of MT Technology
Learning Outcome
Students will be able to prepare data to be used to train MT systems, apply and interpret quality metrics to MT output, and implement basic MT-related algorithms. This outcome is Intellectually Enlarging as it requires students to master the technical "depth" and rigorous logic necessary to handle complex real-world data.
Title
MT Tools and Platforms
Learning Outcome
Students will be able to identify and use available platforms and tools to create MT systems for selected languages. This supports the Character Building aim by requiring the "sustained effort" and technical "integrity" necessary to move from theoretical models to working software.
Title
MT Applications
Learning Outcome
Students will be able to create and incorporate MT systems and/or MT-related functionality in a selected scenario or application. This aligns with the Lifelong Learning and Service aim by training students to use their knowledge to "go forth to serve" a global community through improved accessibility.
Title
Advanced MT algorithms
Learning Outcome
Students will become familiar with and be able to implement and use basic versions of certain more advanced MT algorithms used in current MT research. This fosters Lifelong Learning by developing the ability to "keep pace with the rapidly advancing" state of the art in artificial intelligence.
