Machine Translation, 7.5hp, VT 2016
Course codes: 5LN426 (bachelor) 5LN711 (master)
Course coordinator: Sara Stymne
Examiner: Mats Dahllöf
Teachers: Sara Stymne (SS), Aaron Smith (AS), Christian Hardmeier (CH), Fabienne Cap (FC), Guest lecturers from industry
Reading / Assignments
|F||2016-03-30||10-12||6-K1031||Introduction (SS)||Koehn 1; JM 25.1-2; Hutchins; CFMF|
|F||2016-03-30||14-16||6-K1031||MT evaluation (SS)||Koehn 8; JM 25.9|
|F||2016-04-04||10-12||2-0076||MT in practice (Convertus) - guest lecture|
|L||2016-04-06||10-12||Chomsky||MT in practice (AS)||lab report 1|
|F||2016-04-11||10-12||2-0076||Introduction to SMT (FC)||Koehn Ch 4, Ch 7, KK97|
|L||2016-04-13||10-12||Chomsky||Word-based SMT (SS)||lab report 2|
|L||2016-04-18||10-12||Chomsky||Word-based SMT (SS)||lab report 2|
|F||2016-04-18||14-16||2-1077||Machine translation at Semantix, a translation provider - guest lecture|
|F||2016-04-20||10-12||6-K1031||Parallel Corpora, Alignment (AS)||Koehn 2-4, JT 3-4, KK97, KK99|
|L||2016-04-25||10-12||Chomsky||Parallel corpora & alignment (AS)||lab report 3|
|F||2016-04-27||10-12||2-0076||Phrase-based SMT (FC)||Koehn Ch 5|
|L||2016-05-02||10-12||Chomsky||Phrase-based SMT (AS)||lab report 4|
|F||2016-05-04||10-12||6-K1031||Decoding (CH)||Koehn Ch 6|
|L||2016-05-09||10-12||Chomsky||Phrase-based SMT (AS)||lab report 4|
|F||2016-05-11||10-12||2-0076||Tuning, Factored MT & MT for morphologically rich languages (SS, FC)||Koehn 10.2, 11|
|F||2016-05-16||10-12||2-0076||Document-wide decoding & Neural MT (CH)|
|L||2016-05-18||10-12||Chomsky||Document-wide decoding lab (AS)||oral lab report 5|
|S||2016-05-23||10-12||2-0076||Master projects: Domains, LMs, alignment|
|S||2016-05-25||10-12||6-K1031||Master projects: Reordering, tree-based, compounds|
F=lecture, L=lab, S=seminar.
During the seminars master students will present different topics not presented during the lectures. They are highly recommended for all students, both bachelors and masters. Chomsky = computer lab 9-2043. Turing = computer lab 9-2042
Course EvaluationsA mid-course evaluation was performed on paper during the lecture on April 27. A commented Summary is available.
The final course evaluation will be made at the end and after the course in studentportalen. It will be open May 25 to June 12.
ExaminationRequirements for the Bachelor programme:
- Fulfill all lab requirements and pass all lab assignments (by handing in the requested report or presenting during lab hours, as specified in the lab instructions). Labs are done in pairs of 2 students, and are not graded.
- Fulfill the requirements for the individual assignments. These are done individually and are graded (G/VG/U). Your final grade will determined by this grade.
- Fulfill all lab requirements of labs and pass all lab reports (by handing in the requested report or presenting during lab hours, as specified in the lab instructions). Labs are done in pairs of 2 students, and are not graded.
- Perform a project on a specific topic in MT in a group of 3-4 students, with the folowing two parts:
- Hand in an individual reflection report based on the project work
- Your reflection report and the project work including the seminar presentation and the final report will be graded (VG/G/U), and your final grade will be based on both.
The labs are the same for bachelor and master students.
The labs should be performed in pairs of two students. You can choose a lab partner freely. When you have found a lab partner, sign up for the labs by emailing Sara and Aaron (first_name dot last_name at lingfil dot uu dot se). If you do not find a lab partner on your own, contact us, and we will set you up with someone else. You may not work on your own, and hand in labs on your own, unless there are special circumstances, and you have agreed on this beforehand with Sara and Aaron.
|Lab 1||Evaluation||April 13|
|Lab 2||Word-based SMT||April 26|
|Lab 3||Parallel corpora & alignment||May 3|
|Lab 4||Phrase-based SMT||May 17|
|Lab 5||Document-wide decoding||May 18 during lab (or May 25 with a report)|
If you failed to meet a lab deadline, there will be a final deadline for all labs on June 3.
Individual assignments (bachelor students)
Bachelor students also have to complete 3 individual assignments. The assignments will be graded (U/G/VG), and to receive VG on the course you will need to get a VG on at least two assignments.
|Assignment 1||MT in general and evaluation||May 10|
|Assignment 2a||Select an article||May 16|
|Assignment 2b||Summarize an article||June, 3|
|Assignment 3||Statistical MT||June, 3|
- [Koehn] Philipp Koehn: Statistical Machine Translation,
Cambridge University Press, ISBN-13:
- [KK97] Kevin Knight:
Automating Knowledge Acquisition for Machine Translation
- [KK99] Kevin Knight:
"A Statistical MT Tutorial Workbook", unpublished,
- [Hutchins] John Hutchins: Machine translation: a concise history. To
be published in Computer aided translation: Theory and practice,
ed. Chan Sin Wai. Chinese University of Hong Kong, 2007.
- [CFMF] M.R. Costa-jussà, M. Farrús, J.B. Mariño, J.A.R. Fonollosa. Study and comparison of rule-based and statistical Catalan-Spanish machine translation systems. Computing and Informatics, volume 31, issue 2, pages 245--270. February 2012. ISSN 1335-9150.
- [JM] Speech and Language Processing -
An Introduction to Natural Language Processing,
Computational Linguistics, and Speech Recognition,
by Daniel Jurafsky and James H. Martin
- [JT] Bitext Alignment, Jörg
Tiedemann, Morgan & Claypool Publishers, Synthesis Lectures on
Human Language Technologies, Available as e-book through the
- Bachelor students: A conference paper of your own choice.
- Master students: Literature related to your project topic.
Bokus, adlibris, amazon
Additional material may be added later.