TR EN

AI AND MACHINE TRANSLATION COURSE IDENTIFICATION AND APPLICATION INFORMATION

Code Name of the Course Unit Semester In-Class Hours (T+P) Credit ECTS Credit
ETI216 AI AND MACHINE TRANSLATION 4 3 3 6

WEEKLY COURSE CONTENTS AND STUDY MATERIALS FOR PRELIMINARY & FURTHER STUDY

Week Preparatory Topics(Subjects) Method
1 - Introduction to the class -
2 - The emergence of the idea of automatic translation; mechanical dictionaries and language machines; early ideas about the use of computers in language translation; the development of machine translation research after the Second World War; early automatic translation experiments; the Georgetown–IBM experiment; expectations and limitations in the early period of machine translation. In-class discussion
3 - The basic operating principles of rule-based machine translation; grammatical rules, dictionaries, morphological analysis, and syntactic analysis; the stages of source-language analysis, transfer, and target-text generation; encoding linguistic knowledge into systems. In-class discussion
4 - The direct translation approach; direct transfer of words and structures from the source language to the target language; transfer-based systems; lexical and structural transfer rules; the interlingua approach; comparison of language-pair-dependent and language-independent systems; the Vauquois Triangle. In-class discussion
5 - Dictionary development in rule-based systems; morphological and syntactic ambiguities; idioms, polysemous words, and cultural expressions; the cost of manually creating language rules; problems encountered across different language pairs; analysis of rule-based system outputs. In-class discussion
6 - The emergence of example-based machine translation; translation based on similarity and analogy; reuse of previously translated sentences and expressions; matching, segmenting, and recombining examples; the relationship between translation memory and example-based machine translation. In-class discussion
7 - The concepts of monolingual, bilingual, and parallel corpora; alignment of source and target texts; sentence- and word-level alignment; the impact of corpus size and quality on system performance; training, validation, and test data; data cleaning and data preparation. In-class discussion
8 - The emergence of statistical machine translation; probability-based approaches to translation; translation models and language models; word-based statistical machine translation; word alignment; determining the most probable target sentence for a given source sentence. In-class discussion
9 - Example-based statistical machine translation; translation of groups of words rather than individual words; reordering models; n-gram language models; hierarchical and syntax-based statistical systems; fluency and accuracy problems in statistical systems. In-class discussion
10 - MID-TERM EXAM -
11 - Combining rule-based and statistical methods; hybrid machine translation systems; making use of the strengths of different systems; integration of machine translation into computer-assisted translation tools; combined use of translation memories, terminology databases, and machine translation suggestions. In-class discussion
12 - Sequence-to-sequence models; problems encountered in translating long sentences; the emergence of the attention mechanism; focusing on different parts of the source sentence; use of contextual information; fluency, accuracy, omission, and addition errors in neural machine translation. In-class discussion
13 - The emergence of Transformer architecture; the self-attention mechanism; encoder and decoder layers; parallel processing and the use of large datasets; multilingual neural machine translation; zero-shot translation; domain adaptation; low-resource languages; customization of neural machine translation systems. In-class discussion
14 - Basic working principles of large language models; pretrained language models; direct translation with generative AI; providing context, purpose, target audience, terminology, and style information; translation-oriented prompt design; step-by-step prompting; comparison of machine translation engines and large language models; risks of hallucinated information and meaning shifts in AI-generated outputs. In-class discussion
15 - Comparison of human translation, machine translation, and AI-assisted translation; light and full post-editing of machine translation and AI-generated translation; error classification and quality assessment; the translator’s ultimate responsibility; confidentiality, data security, copyright, algorithmic bias, and transparency; disclosure of AI use; the changing professional roles of translators and future human–AI workflows. In-class discussion
16 - FINAL EXAM -
17 - FINAL EXAM -