Infopaket

Machine Translation Systems

ДВС. 1.04. 03

Вибіркова дисципліна для ОП

First

2022/2023

7 Semester

2

PLO 15. To conduct linguistic analysis (phonetic, morphemic, syntax, word formation, morphological, semantic), literary criticism, and translation studies analysis of texts of different genres and styles, as well as computer modeling of analysis and synthesis of linguistic objects and phenomena. PLO 16. To know and understand the major notions, theories, and conceptions of structural, mathematical, and computational linguistics, and to be able to use them in professional activity. PLO 17. To collect, analyze, systematize (to create system-oriented and text-oriented linguistic database) and interpret language facts and speech, and to use them for solving complicated tasks and problems of automatic linguistic analysis in specialized spheres of professional and/or learning activity. PLO 21, 25.

Full-time form

1) Successful mastering of the complex disciplines of Theoretical Linguistics, Applied Linguistics, Informatics, Mathematical Foundations of the specialty, disciplines Basics of automatic translation; 2) Knowledge of the theoretical foundations of general linguistics, mathematical and computer linguistics, machine translation, computer science, and programming; 3) Mastery of linguistic analysis techniques at all language levels; basics of corpus technologies, skills of mathematical and computer modeling of linguistic phenomena; mathematical basics of computer technology; 4) Ability to perform expert and automatic evaluation of machine translation results, corpus markup at different levels, and use a programming language to solve problems of input, output, and processing of text information.

The discipline "Machine Translation Systems" is a logical continuation of the course "Basics of automatic translation". Within the scope of this discipline, implementation of methods and algorithms for automatic translation from one language to another, evaluation of parameters of translation models, and analysis of translation quality are considered in depth; attention is paid to the peculiarities of subject areas, styles, modalities, and computing resources, which affects the choice of approaches, their implementation and the design of machine translation systems in general.

Bapna, A., & Firat, O. (2019). Exploring massively multilingual, massive neural machine translation. Google AI Blog, October, 11. Dabre, R., Chu, C., & Kunchukuttan, A. (2020). A survey of multilingual neural machine translation. ACM Computing Surveys (CSUR), 53(5), 1-38. Fan, A., Bhosale, S., Schwenk, H., Ma, Z., El-Kishky, A., Goyal, S., ... & Joulin, A. (2021). Beyond english-centric multilingual machine translation. The Journal of Machine Learning Research, 22(1), 4839-4886. Forcada, M. L., Bonev, B. I., Rojas, S. O., Ortiz, J. P., Sánchez, G. R., Martínez, F. S., ... & Tyers, F. M. (2007). Documentation of the open-source shallow-transfer machine translation platform Apertium.

Lectures, practical classes, and independent work. Types of work: The answer is in the practical lesson; Lab; Report-analysis of scientific literature; Tests.

During the semester, practical classes are held on relevant topics, at which evaluation is carried out according to the types of work. The final assessment is conducted in the form of a comprehensive exam: ● the maximum number of points on the exam is 40 points, the minimum number of points (positive score) that are added to the semester score is 24 points (60% of the maximum number of points assigned to the exam); ● the exam is conducted in the form of a written work and involves the performance of three tasks: the answer to two theoretical questions, each of which is evaluated with a maximum of 10 points (25% of the final score), as well as the performance of an individual practical task, which is evaluated with a maximum of 20 points (50% from the final assessment). The semester final score is formed by the points received by the student during the semester and the points received on the exam: 60 points (60%) - semester control and 40 points (40%) - exam).

Ukrainіаn