ﺩﺍﻧﺸگﺎﻩ ﺻﻨﻌﺘﻲ ﺍﻣﻴﺮ کﺒﻴﺮ ﺩﺍﻧﺸکﺪﻩ ﻣﻬﻨﺪﺳﻲ کﺎﻣپﻴﻮﺗﺮ

ﺩﺍﻧﺸگﺎﻩ ﺻﻨﻌﺘﻲ ﺍﻣﻴﺮ کﺒﻴﺮ ﺩﺍﻧﺸکﺪﻩ ﻣﻬﻨﺪﺳﻲ کﺎﻣپﻴﻮﺗﺮ ﻭ ﻓﻨﺎﻭﺭﻱ ﺍﻃﻼﻋﺎﺕ INTERLINGUAL MACHINE TRANSLATION ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒﻴﻌی ﺍﺳﺘﺎﺩ ﺩﺭﺱ: ﺩکﺘﺮ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی- 13013168 1

ﻣﺮﻭﺭی کﻮﺗﺎﻩ ﺑﺮ ﺗﺮﺟﻤﻪ ﻣﺎﺷیﻨی چیﺴﺖ ؟ Automated system Analyzes text from Source Language (SL) Produces “equivalent” text in Target Language (TL) Ideally without human intervention Source Language Target Language ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

ﻣﺮﻭﺭی کﻮﺗﺎﻩ ﺑﺮ ﺗﺮﺟﻤﻪ ﻣﺎﺷیﻨی ﺭﻭﺵ ﻫﺎی ﺍﺻﻠی ﺗﺮﺟﻤﻪ ﻣﺎﺷیﻨی Direct Transfer Interlingual ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

Direct ﺍیﻦ ﺭﻭﺵ ﺩﺭﻭﺍﻗﻊ ﺗﺮﺟﻤﻪ ﻟﻐﺖ ﺑﻪ ﻟﻐﺖ ﻣی ﺑﺎﺷﺪ. Transfer ﺯﺑﺎﻥ ﻣﺒﺪﺍ ﺑﻪ یکی ﺍﺯ ﺑﺎﺯﻧﻤﺎیی ﻫﺎی Syntax یﺎ Semantic ﺗﺤﻠیﻞ ﺷﺪﻩ ﻭ پﺲ ﺍﺯ ﺍیﻦ ﺑﺎﺯﻧﻤﺎیی ﺯﺑﺎﻥ ﻣﺒﺪﺍ ﺑﻪ ﺑﺎﺯﻧﻤﺎیی ﻣﻨﺎﺳﺐ ﺯﺑﺎﻥ ﻣﻘﺼﺪ ﺗﺒﺪیﻞ ﺷﺪﻩ ﻭ ﺩﺭ ﻧﻬﺎیﺖ ﺟﻤﻼﺕ ﺯﺑﺎﻥ ﻣﻘﺼﺪ ﺍﺯ ﺍیﻦ ﺑﺎﺯﻧﻤﺎیی ﺗﻮﻟیﺪ ﻣی ﺷﻮﻧﺪ. Interlingual ﺟﻤﻼﺕ ﺯﺑﺎﻥ ﻣﺒﺪﺍ ﺑﻪ یک ﺑﺎﺯﻧﻤﺎیی ﻣﻔﻬﻮﻣی ﺳﺮﺍﺳﺮی کﻪ ﺑﻪ آﻦ IL گﻔﺘﻪ ﻣی ﺷﻮﺩ ،ﺗﺒﺪیﻞ ﺷﺪﻩ ﻭ ﺟﻤﻼﺕ ﺯﺑﺎﻥ ﻣﻘﺼﺪ ﺍﺯﺗﺒﺪیﻞ آﻦ ﺑﺪﺳﺖ ﻣی آیﺪ. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

Direct Transfer Interlingual ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

Direct Transfer Interlingual

Three main methodologies for Machine Translation Direct Transfer Interlingual

ﺑﺨﺶ ﺩﻭﻡ کﺘﺎﺏ ﻣﺴﺎﺋﻞ ﻣﻄﺮﺡ ﺩﺭﺳﺎﺧﺖ ﺳیﺴﺘﻢ ﻫﺎی Large Scale ﻭ General Purpose پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی , Nonuniform knowledge represantation knowledge acquisition ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 8 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ §Uniform §Automatic

Uniform and Nonuniform knowledge represantation § Uniform ﺩﺭ ﺭﻭﺵ یکﻨﻮﺍﺧﺖ،ﺑﺮﺍی ﺗﻤﺎﻡ ﻓﻌﺎﻟیﺖ ﻫﺎ ) (Task ﻭ ﻣﻮﻟﻔﻪ ﻫﺎ ﺍﺯ یک ﺯﺑﺎﻥ ﺑﺎﺯﻧﻤﺎیی ﺍﺳﺘﻔﺎﺩﻩ ﺷﺪﻩ ﺍﺳﺖ. ﺑﺮﺍی ﻣﺜﺎﻝ : ﺍﺳﺘﻔﺎﺩﻩ ﺍﺯ ﻣﻨﻄﻖ ﻣﺮﺗﺒﻪ ﺍﻭﻝ ﺑﺮﺍی ﺗﻤﺎﻡ ﺑﺨﺶ ﻫﺎ. ﺑﺰﺭگﺘﺮیﻦ ﻣﺸکﻞ ﺍیﻦ ﺭﻭﺵ : ﺍیﻨکﻪ ﺳیﺴﺘﻢ ﻫﺎی ﺑﺎﺯﻧﻤﺎیی ﺩﺍﻧﺶ ﺩﺭ ﺩﺳﺘﺮﺱ ﻗﺎﺑﻠیﺖ ﺑﺎﺯﻧﻤﺎیی ﺗﻤﺎﻡ ﺧﺼﻮﺻیﺎﺕ ﺯﺑﺎﻥ ﻃﺒیﻌی ﺭﺍ ﻧﺪﺍﺭﻧﺪ. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 9 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

Uniform and Nonuniform knowledge represantation § Nonuniform ﺩﺭ ﺭﻭﺵ ﻏیﺮیکﻨﻮﺍﺧﺖ، ﺍﺯ ﺯﺑﺎﻥ ﻫﺎی ﺑﺎﺯﻧﻤﺎیی ﻣﺨﺘﻠﻔی ﺑﺮﺍی ﺑﺎﺯﻧﻤﺎیی ﺩﺍﻧﺶ ﺑﺮﺍی ﻓﻌﺎﻟیﺖ ﻫﺎ ﻭ ﻣﻮﻟﻔﻪ ﻫﺎی ﻣﺨﺘﻠﻒ ﺍﺳﺘﻔﺎﺩﻩ ﻣی کﻨﺪ. ﻣﻬﻤﺘﺮیﻦ ﻣﺸکﻞ ﺍیﻦ ﺭﻭﺵ : ﻟﺰﻭﻡ ﺗﺮﺟﻤﻪ ﺑیﻦ ﺑﺎﺯﻧﻤﺎیی ﻫﺎ ﻣﺘﻔﺎﻭﺕ ﺑﺮﺍی ﺗﺒﺪیﻞ ﻭﺗﺮکیﺐ ﺩﺍﻧﺶ. ﺍیﻦ ﺍﻣﺮ ﺩﺭ ﺳیﺴﺘﻢ ﻫﺎی ﺑﺰﺭگ ﻭ ﺑﻮیژﻪ ﺑﺮﺍی ﺩﺍﺩﻩ ﻫﺎی ﻭﺍﻗﻌی ﺑﺴیﺎﺭ پﺮﻫﺰیﻨﻪ ﻭ پیچیﺪﻩ ﻣی ﺑﺎﺷﺪ. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 01 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

A Multi-Level Approach to Interlingual MT Defining the Interface between Representational Languages Bonnie J. Dorr and Clare R. Voss Department of Computer science University of Maryland ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ 11

کﻠیﺖ ﻣﻮﺿﻮﻉ ﻣﻘﺎﻟﻪ • یک ﻃﺮﺍﺣی چﻨﺪ ﻻیﻪ ﺍی ﺑﺮﺍی یک ﺳیﺴﺘﻢ ﺗﺮﺟﻤﻪ ﻣﺎﺷیﻨی ﺷﺮﺡ ﺩﺍﺩﻩ ﻣی ﺷﻮﺩ. • یک ﺳیﺴﺘﻢ ﻏیﺮ یکﻨﻮﺍﺧﺖ کﻪ ﺑﺮﺍی ﺗﻮﺻیﻒ ﺩﺍﻧﺶ ﻫﺎی ﻣﺘﻔﺎﻭﺕ ﺍﺯ ﺯﺑﺎﻥ ﻫﺎی ﺑﺎﺯﻧﻤﺎیی ﻣﺨﺘﻠﻒ ﺍﺳﺘﻔﺎﺩﻩ ﺷﺪﻩ ﺍﺳﺖ. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 21 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

ﺑﺮﺍی ﺗﻮﻟیﺪ یک ﺗﺮﺟﻤﻪ ﺧﻮﺏ ﺍﺯیک ﺟﻤﻠﻪ ،ﺑﺎیﺪ یک ﺳیﺴﺘﻢ ﺗﺮﺟﻤﻪ ﻣﺎﺷیﻨی ﺑﻪ چﻨﺪ . ﺭﻭﺵ ﺑﺎﺯﻧﻤﺎیی ﺩﺳﺘﺮﺳی ﺩﺍﺷﺘﻪ ﺑﺎﺷﺪ Lexical for lexicon-based information Syntactic for defining phrase structure interlingual (or IL) for sentence interpretation knowledge representational (or KR) for filtering out interpretations that are incompatible with facts in the MT system's knowledge base. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ 13

This paper examines the interface between the interlingua and other representation types in an interlingual MT system. multi-level : syntactic, IL and KR And non-uniform approach : in which distinct representational languages are used for different types of knowledge. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ 14

SL Input TL Output SL Syntactic Analysis SL Lexicon TL Syntactic Analysis IL Composition and Decomposition TL Lexicon KR Filtering and Inference 208 (،ﺻﻔﺤﻪ Interlingual Machine Translation) 6 ﺷکﻞ 1 ،کﺘﺎﺏ ، ﻓﺼﻞ ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ 15

ﻓﺎﺯﻫﺎی پﺮﺩﺍﺯﺵ ﻫﻤﺎﻧﻄﻮﺭ کﻪ ﺩﺭ ﺷکﻞ 1 ﻣﺸﺎﻫﺪﻩ ﻣی ﺷﻮﺩ، ﺩﺭ ﺍیﻦ ﻣﺪﻝ پیﺸﻨﻬﺎﺩی 3 ﻓﺎﺯ پﺮﺩﺍﺯﺵ . ﺩﺍﺭیﻢ 1 _ Analysis/synthesis phase : in which a source-language (SL) sentence is parsed into a syntactic structure. 2 _A composition/decomposition phase : A SL syntactic structure is composed into an IL representation or an IL representation is decomposed into a TL syntactic structure and lexical items. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ 16

3_ KR phase : checks the IL representations • filtering out incompatible forms with known facts • Coercing or augmenting IL forms with logically inferred knowledge in order to resolve an incomplete IL composition. • ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ 17

کﺎﺭﻫﺎی گﺬﺷﺘﻪ Voss ﻭ Dorr ﺩﺭ ﺳﺎﻝ 3991 ﻣی گﻮیﺪ : کﻤﺒﻮﺩی کﻪ ﺩﺭ ﺯﻣیﻨﻪ ﺗﺤﻘیﻘﺎﺕ ﻭ ﺳﺎﺧﺖ IL ﻭﺟﻮﺩ ﺩﺍﺭﺩ،ﺍیﻦ ﺍﺳﺖ کﻪ: ﺍﺗﻔﺎﻕ ﻧﻈﺮی ﺑﺮ ﺍیﻨکﻪ interlingua چیﺴﺖ ﻭ چگﻮﻧﻪ ﺗﻌﺮیﻒ ﻣی ﺷﻮﺩ ، ﻭﺟﻮﺩ ﻧﺪﺍﺭﺩ. ﺑﺮﺍی ﻣﺜﺎﻝ : (Rosetta, 1994) used an interlingua based on . Montague-grammar Mikrokosmos (1994) developed based on their ) own Text Meaning Representation (TMR . language ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 81 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

کﺎﺭﻫﺎی گﺬﺷﺘﻪ Verkuyl (1994) : a "layered" interlingua in two layer • Discourse Representation Structures • one level a Lexical Conceptual Structures ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ 19

ﺩﺭ ﺍیﻦ ﻣﻘﺎﻟﻪ PRINCIRAN : • Interpretation and representation of natural language sentences. ﺩﺭﻭﺍﻗﻊ ﺍیﻦ ﺳیﺴﺘﻢ ﺑﺮﺍی ﺳﺎﺧﺖ یک ﺳیﺴﺘﻢ ﺑﺰﺭگ 3 ﺳیﺴﺘﻢ ﺭﺍ ﺗﺮکیﺐ ﻣی . کﻨﺪ syntactic processing design of PRINCIPAR (Dorr, Lin, Lee, and Suh (1995)) syntax-IL interface UNITRAN (Dorr, 1993) IL-KR interface from the LEXITRAN (Dorr and Voss, 1993) ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ 20

ﻣﺜﺎﻟی ﺑﺮﺍی ﺑیﺎﻥ، چﺮﺍیی ﺍﻧﺘﺨﺎﺏ یک ﺭﻭﺵ ﻏیﺮ یکﻨﻮﺍﺧﺖ : German sentence “ “ Die Kirche liegt im S"uden der Stadt ﺍیﻦ ﺟﻤﻠﻪ ﻣی ﺗﻮﺍﻧﺪ ﻫﺮ یک ﺍﺯ ﺩﻭ ﺗﻔﺴیﺮ ﺯیﺮ ﺭﺍ ﺩﺍﺷﺘﻪ ﺑﺎﺷﺪ. The church lies in the south of the city ) (southern part of the city The church lies to the south of the city ) (south of the city , outside the city ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 12 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

ﻣﺜﺎﻟی ﺑﺮﺍی ﺑیﺎﻥ، چﺮﺍیی ﺍﻧﺘﺨﺎﺏ یک ﺭﻭﺵ ﻏیﺮ یکﻨﻮﺍﺧﺖ ﺍیﻦ کﺎﻣﻼ ﻭﺍﺿﺢ ﺍﺳﺖ کﻪ ﺟﻤﻠﻪ ﻣﺰﺑﻮﺭ ﺩﺭ ﺯﺑﺎﻥ آﻠﻤﺎﻧی ﻫیچ ﺍﺑﻬﺎﻣی ﻧﺪﺍﺭﺩ، ﺍﻣﺎ یک ﺳیﺴﺘﻢ ﺗﺮﺟﻤﻪ ﻣﺎﺷیﻨی ﺑﺎیﺪ ﺑﺪﺍﻧﺪ کﻪ ﺟﻤﻠﻪ im S"uden der Stadt ﺑﻪ ﺩﻭ ﺷکﻞ ﻣﺘﻤﺎیﺰ ﺑﺎﺯﻧﻤﺎیی ﻣی ﺷﻮﺩ. 1 - south-and-internal 2 - south-and-external ﺩﺭ ﻭﺍﻗﻊ ﺍیﻦ کﺎﺭیک KR filtering function ﺑﻮﺩﻩ ﻭ ﺟﺪﺍ ﺍﺯ knowledge lexical ﻭ یﺎ Interlingua ﻣی ﺑﺎﺷﺪ. ﺍیﻦ ﻣﻬﻤﺘﺮیﻦ ﺑﺨﺶ کﺎﺭ ﺍیﻦ پﺮﻭژﻪ ﺍﺳﺖ. ﺍیﻦ کﺎﺭ یک ﺷیﻮﻩ ﺧﺎﺹ ﺑﺮﺍی KR ﻣی ﺑﺎﺷﺪ، ﻧﻪ . IL ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 22 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

ﻣﺜﺎﻟی ﺑﺮﺍی ﺑیﺎﻥ، چﺮﺍیی ﺍﻧﺘﺨﺎﺏ یک ﺭﻭﺵ ﻏیﺮ یکﻨﻮﺍﺧﺖ Using Default knowledge in the KR Mountains are physical entities, typically distinct and external to cities System chooses second translation The mountain lies to the south of the city Using specific facts in the KR A particular mountain is in the city System overrides default knowledge and chooses first translation The mountain lies in the south of the city

ﻓﺮﺿیﺎﺕ پﺮﻭژﻪ 1 پﺮﺩﺍﺯﺵ ﻫﺎ ﺗﻨﻬﺎ ﺩﺭ sentence-level ﺑﻮﺩﻩ ﻭ آﻨﺎﻟیﺰ) (discourse ﻣﻮﺭﺩ ﻧﻈﺮ ﻧﻤی ﺑﺎﺷﺪ. 2 ﻭﺭﻭﺩی ﺳیﺴﺘﻢ، ﺧﺮﻭﺟی ﺳیﺴﺘﻢ PRINCIPAR ﺍﺳﺖ. PRINCIPAR پﺎﺭﺳﺮﺍﺳﺘﻔﺎﺩﻩ ﺳﺎﺧﺘﻪ ﺷﺪﻩ ﺗﻮﺳﻂ Lin،Dorr ﻭ Lee ﺩﺭ ﺳﺎﻝ 5991 ﺍﺳﺖ. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 42 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

ﻓﺮﺿیﺎﺕ پﺮﻭژﻪ 3 ﺯﺑﺎﻥ ﻣﺒﺪﺍ ﺑﻪ ﺻﻮﺭﺕ ﻣﺠﻤﻮﻋﻪ ﺍی ﺍﺯ Parse Tree ﺗﺒﺪیﻞ ﺷﺪﻩ، کﻪ ﺩﺭ ﺍیﻦ ﻣﺠﻤﻮﻋﻪ ﺗﻤﺎﻡ ﺑﺎﺯﻧﻤﺎیی ﻫﺎی ﻧﺤﻮی ﻣﻤکﻦ ﺍﺯ ﺟﻤﻠﻪ ﻣﺒﺪﺍ ﻓﺮﺍﻫﻢ ﺷﺪﻩ ﺍﺳﺖ. 4 آﻨﺎﻟیﺰ Phrase Structure ﻭ ﺍیﺠﺎﺩ ﺑﻬﺘﺮیﻦ ﺗﻔﺴیﺮ ﺑیﻦ ﺯﺑﺎﻧی ﺑﺮﺍی ﺗﻮﻟیﺪ ﺯﺑﺎﻥ ﻣﻘﺼﺪ، ﻭﻇیﻔﻪ ﻣﻮﻟﻔﻪ ﻫﺎی IL ﻭ KR ﺍﺳﺖ. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 52 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

Defined interfaces between three knowledge levels 1 -Relates IL representations to corresponding syntactic forms by means of lexical entries. 2 - Checks the IL representations in the KR, filtering out those forms incompatible with known facts. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ 26

Defining the Problem Space: Translation Mismatches ﺩﺭ ﺍیﻦ ﺑﺨﺶ ﺣﻞ ﻣﺴﺌﻠﻪ ﻋﺪﻡ ﺗﻄﺎﺑﻖ ﺩﺭ ﺗﺮﺟﻤﻪ ﻫﺎ ﻣﻮﺭﺩ ﻧﻈﺮ ﺍﺳﺖ. ﺩﺭ ﺍیﻦ ﺯﻣیﻨﻪ ﺑﻪ ﺑیﺎﻥ ﺩﻭ گﺮﻭﻩ ﺍﺯ ﺍﺧﺘﻼﻑ ﻫﺎیی کﻪ ﺑیﻦ ﺟﻤﻼﺕ ﻣﺒﺪﺍ ﻭ ﻣﻘﺼﺪ ﻣی ﺗﻮﺍﻧﺪ ﻭﺟﻮﺩ ﺩﺍﺷﺘﻪ ﺑﺎﺷﺪ، ﺑیﺎﻥ ﻣی کﻨیﻢ. translation divergences ﻣﻔﻬﻮﻡ ﺟﻤﻠﻪ ﺩﺭ ﺯﺑﺎﻥ ﻣﺒﺪﺍ کﺎﻣﻼ ﻣﻨﺘﻘﻞ ﺷﺪﻩ ﺍﻣﺎ ﺳﺎﺧﺘﺎﺭ ﺟﻤﻼﺕ ﺩﺭ ﺩﻭ ﺯﺑﺎﻥ ﻣﺘﻔﺎﻭﺕ ﺍﺳﺖ. translation mismatches ﻣﻔﻬﻮﻡ ﻣﻨﺘﻘﻞ ﺷﺪﻩ ﺩﺭ ﺩﻭ ﺯﺑﺎﻥ ﺑﺎ ﻫﻢ ﻣﺘﻔﺎﻭﺕ ﺍﺳﺖ. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 72 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

translation divergences Conflational Divergence Translation of two or more words in one language into one word in another language To kick Dar una patada (Give a kick)

translation divergences ﻋﺪﻡ ﺗﻄﺎﺑﻖ ﺑﻪ ﺩﻟیﻞ ﺗﻔﺎﻭﺕ ﻫﺎی ﺳﺎﺧﺘﺎﺭی ﺯﺑﺎﻥ ﻫﺎ Realization of verb arguments in different syntactic configurations in different languages To enter the house Entrar en la casa (Enter in the house)

Lexical Mismach

ﺗﻌییﻦ ﻣﺤﺪﻭﺩﻩ ﺩﺭ ﺍیﻦ پژﻮﻩ ﺑیﺸﺘﺮیﻦ ﺗﺎکیﺪ ﺑﺮ Spatial Expression ﻭ ﺑﻮیژﻪ ﺑﺮ Spatial Verb ﺍﺳﺖ کﻪ ﺑﻪ آﻨﻬﺎ Spatial Predicates گﻔﺘﻪ ﻣی ﺷﻮﺩ. Spatial Predicate گﺰﺍﺭﻩ ﻫﺎیی کﻪ ﺑﺮﺍی ﺗﻮﺻیﻒ ﺍﺯﺗﺒﺎﻁ ﺑیﻦ ﺍﺷیﺎﺀ ﻓیﺰیکی ﺩﺭ ﻓﻀﺎی ﺳﻪ ﺑﻌﺪی ﺑﻪ کﺎﺭ ﻣی ﺭﻭﻧﺪ. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 13 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

ﺭﺍﻩ ﺣﻞ ﻫﺎی ﺩﻧﺒﺎﻝ ﺷﺪﻩ ﺩﺭ پژﻮﻫﺶ ﻫﺎی پیﺸیﻦ • ﺣﻞ ﻣﺸکﻞ : divergences ﺗﻐییﺮ ﻭ ﺑﻬﺒﻮﺩ ﺑﺨﺸیﺪﻥ ﻣکﺎﻧیﺰﻡ ﻫﺎی ﺗﺒﺪیﻞ، ﺑﺎﺯﻧﻤﺎیی IL ﺑﻪ ﺳﺎﺧﺘﺎﺭ ﺯﺑﺎﻥ ﻣﻘﺼﺪ. • ﺣﻞ ﻣﺸکﻞ : mismatch ﺗﺎکیﺪ ﺑیﺸﺘﺮ ﺑﺮﺍﺭﺍیﻪ ﺟﺰﺋیﺎﺕ ﺑﺎﺯﻧﻤﺎیی ﻣﻔﻬﻮﻣی ﺩﺭ ﺳﺎﺧﺖ IL ﺍﺳﺖ. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 23 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

ﺭﺍﻩ ﺣﻞ ﺑیﺎﻥ ﺷﺪﻩ ﺩﺭ ﺍیﻦ پژﻮﻫﺶ ﺣﻞ ﻫﺮ ﺩﻭ ﻣﺸکﻞ ﻣﻄﺮﺡ ﺷﺪﻩ ﺩﺭ یک ﺳیﺴﺘﻢ. ﺣﻞ ﻣﺸکﻞ : mismatch ﺑﺎ ﺩﺳﺘﺮﺳی ﺑﻪ KR ﻭﺍﺳﺘﻔﺎﺩﻩ ﺍﺯ ﺩﺍﻧﺸی کﻪ ﺑﻄﻮﺭ ﻣﻨﻄﻘی ﻗﺎﺑﻞ ﺍﺳﺘﻨﺘﺎﺝ ﺍﺳﺖ. ﺣﻞ ﻣﺸکﻞ : divergences ﺳﺎﺧﺖ ﺗﻌﺪﺍﺩ کﺎﻓی ﺳﺎﺧﺘﺎﺭ ﺩﺭ IL ﻭ ﺍﺳﺘﻔﺎﺩﻩ ﺍﺯ آﻨﻬﺎ)ﺍﺯ ﺑیﻦ ﺑﺮﺩﻥ ﺗﻔﺎﻭﺕ ﻫﺎی ﺳﺎﺧﺘﺎﺭی(. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 33 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

ﺗﻌﺮیﻒ یک Interlingua ﺑﺮﺍی ﺗﻌﺮیﻒ Interlingua ﺍﺯ ﺳﻪ ﻣﻨﺒﻊ ﺍﺳﺘﻔﺎﺩﻩ ﺷﺪﻩ ﺍﺳﺖ. ) 1 - Lexical Conceptual Structure(LCS ) 2 – Lexical Semantic Template(LST 3 – Semantic Classification Scheme ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 43 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

Lexical Conceptual Structure(LCS) By Jackendoff (1983 , 1990) Three independent subsystem • Fields • Conceptual constituents • Boundedness and aggregation property . ﺗﻨﻬﺎ ﺍﺯ ﺩﻭ ﺯیﺮ ﺳیﺴﺘﻢ ﺍﺑﺘﺪﺍ ﺩﺭ ﺍیﻦ پﺮﻭژﻪ ﺍﺳﺘﻔﺎﺩﻩ ﺷﺪﻩ ﺍﺳﺖ ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ 35

) Lexical Conceptual Structure(LCS Fields ﺍیﻦ ﻓیﻠﺪ ﻫﺎ ﺗﻮﺳﻂ ﻣﺸﺎﻫﺪﺍﺕ ﺗﺨﺼﺼی ﺩﺭ ﺗﻘﺎﺭﻥ ﻫﺎی ﻟﻐﻮی ﺳﺎﺧﺘﻪ ﺷﺪﻩ ﺍﻧﺪ، ﻣی ﺗﻮﺍﻧﻨﺪ ﺍﺯ ﻗﺒیﻞ (, Loc(ational), Temp(oral), Poss(essional ) Ident(ificational), Perc(eptual ﺑﺎﺷﻨﺪ. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 63 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

) Lexical Conceptual Structure(LCS Conceptual constituents گﺰﺍﺭﻩ ﻫﺎی ﺍﻭﻟیﻪ GO, STAY, BE, GO-EXT, and ORIENT • آﺮگﻮﻣﺎﻥ ﻫﺎ ﻭ ﺗﻮﺻیﻒ کﻨﻨﺪﻩ ﻫﺎی گﺰﺍﺭﻩ ﻫﺎی ﺍﻭﻟیﻪ • Type یﺎ Antological Type گﺰﺍﺭﻩ ﻫﺎی ﺍﻭﻟیﻪ Thing, State, Event, Place, Path, and Property ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 73 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

) Lexical Conceptual Structure(LCS یک ﻣﺜﺎﻝ ﺍﺯ ﺍیﻨکﻪ ﺑﺒیﻨیﻢ چگﻮﻧﻪ گﺰﺍﺭﻩ ﺍﻭﻟیﻪ Go ﺑﺎ ﻧﻮﻉ Event ﺑﺮﺍی ﺑﺎﺯﻧﻤﺎیی ﻣﻔﻬﻮﻡ یک ﺟﻤﻠﻪ ﺑﻪ کﺎﺭ ﻣی ﺭﻭﺩ. ﺍیﻦ ﺑﺎﺯﻧﻤﺎیی ﺑﻪ ﺍیﻦ ﻣﻌﻨﺎ ﺍﺳﺖ کﻪ "The ball went locationally ". toward Beth ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 83 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

) Lexical Conceptual Structure(LCS ﺩﺭ ﺍیﻨﺠﺎ یک ﺷکﻞ ﺑﺎﺯﻧﻤﺎیی ﺑﺮﺍی ﺟﻤﻠﻪ ﺯیﺮ ﺩﺍﺭیﻢ. ” “John jogged to school ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 93 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

) Lexical Conceptual Structure(LCS ﺑﺎﺯﻧﻤﺎیی ﺑیﺎﻥ ﺷﺪﻩ ، ﺗﻨﻬﺎ ﻣﻔﻬﻮﻡ ﺑﻪ ﻣﺪﺭﺳﻪ ﺭﻓﺘﻦ ﺭﺍ ﻣی ﺭﺳﺎﻧﺪ ﻭ ﺩﺭ ﺍیﻦ ﺑﺎﺯﻧﻤﺎیی ﺗﻔﺎﻭﺗی ﺑیﻦ Walk ، Run ﻭ یﺎ Jog ﺑیﺎﻥ ﻧﺸﺪﻩ ﺍﺳﺖ. ﺑﻨﺎﺑﺮﺍیﻦ ﺷکﻞ کﺎﻣﻞ ﺍیﻦ ﺑﺎﺯﻧﻤﺎیی ﺑﻪ ﺻﻮﺭﺕ ﺯیﺮ ﺍﺳﺖ. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 04 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

Lexical Conceptual Structure(LCS) LCS ﻧﺘیﺠﻪ گیﺮی ﺩﺭ ﺑﺤﺚ • The LCS approach views semantic representation as a subset of conceptual Structure • This representation abstracts away from syntax just far enough to enable language-independent encoding. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ 41

) Lexical Semantic Template(LST ﺍیﻦ ﺍﺑﺰﺍﺭ ﺍﻓﻌﺎﻝ ﺭﺍ ﺑﻪ ﺩﻭ ﺑﺨﺶ Predicate Structure ﻭ Nonpredicate constant ﺗﺠﺰیﻪ ﻣی کﻨﺪ. کﻪ ﺑﺨﺶ ﺩﻭﻡ ﺭﺍ ﺑﻪ ﺷکﻞ >ﺷکﻞ ﺛﺎﺑﺖ ﻓﻌﻞ< ﻧﻤﺎیﺶ ﺩﺍﺩﻩ ﻣی ﺷﻮﺩ. یک ﻓﻌﻞ ﺑﺎ چﻨﺪ ﻣﻌﻨﺎ ﺩﺍﺭﺍی یک ﺛﺎﺑﺖ ﻭ چﻨﺪ ﺳﺎﺧﺘﺎﺭ گﺰﺍﺭﻩ ﺍی ﻣﺘﻔﺎﻭﺕ ﻣی ﺑﺎﺷﺪ. ﺟﻤﻼﺕ ﺑﺮ ﻋﻬﺪﻩ ﺳﺎﺧﺘﺎﺭ گﺰﺍﺭﻩ ﺍی ﻭ ﺗﺸﺨیﺺ ﻣﻌﻨﺎ ﺩﺭ ﻫﺮ یک ﺍﺯ ﻣﺮﺑﻮﻃﻪ ﺍﺳﺖ. ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 24 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

Lexical Semantic Template(LST) : ﻣﺜﺎﻝ The soldiers marched to the barracks. The soldiers marched clear of the falling rocks. The soldiers marched the soles of their boots flat. The general marched the soldiers to the barracks. <march> ﺩﺭ ﺍیﻦ ﻣﺜﺎﻝ ﺗﻨﻬﺎ یک ﺷکﻞ ﺛﺎﺑﺖ ﺑﺮﺍی ﻓﻌﻞ ، ﺑﻪ ﺷکﻞ . ﺩﺍﺭیﻢ، ﺍﻣﺎ ﻣﻔﻬﻮﻡ ﻣﺘﻔﺎﻭﺕ ﺩﺭﻫﺮ ﺟﻤﻠﻪ ﺑﺮ ﻋﻬﺪﻩ ﺳﺎﺧﺘﺎﺭ گﺰﺍﺭﻩ ﺍی ﺍﺳﺖ ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ 43

) Lexical Semantic Template(LST ﺍﻣکﺎﻥ ﺩیگﺮ ﺍیﻦ ﺍﺑﺰﺍﺭ ﺩﺍﺷﺘﻦ چﻨﺪ Constant ﻣﺘﻔﺎﻭﺕ ﺑﺎ یک ﺳﺎﺧﺘﺎﺭ گﺰﺍﺭﻩ ﺍی یکﺴﺎﻥ ﺍﺳﺖ. ﻫﺮ ﺗﺮکیﺐ ﺍﺯ ﺍیﻨﻬﺎ ﺑﺼﻮﺭﺕ یک ﻓﻌﻞ ﺟﺪﺍ ﺩﺭ ﺟﻤﻠﻪ ﺍﺻﻠی ﺷﻨﺎﺧﺘﻪ ﻣی ﺷﻮﺩ. ﻣﺜﺎﻝ . They funneled the mixture into the jar . They ladled the mixture into the jar . They spooned the mixture into the jar ﺗﻨﻬﺎ ﺗﻔﺎﻭﺕ ﺩﺭ ﺍﺑﺰﺍﺭ کﺎﺭ ﺍﺳﺖ. > <LADLE> ، <FUNNEL ﻭ>. <SPOON ﺩﺭﺱ پﺮﺩﺍﺯﺵ ﺯﺑﺎﻥ ﻃﺒیﻌی، ﺍﺳﺘﺎﺩ: ﺩکﺘﺮ ﺍﺣﻤﺪ ﻋﺒﺪﺍﻟﻠﻪ ﺯﺍﺩﻩ ﺗﻮﺳﻂ: ﺍﺳﻤﺎﻋیﻞ ﺭﺿﺎیی 44 آﺰﻣﺎیﺸگﺎﻩ ﺳیﺴﺘﻢ ﻫﺎی ﻫﻮﺷﻤﻨﺪ

References Journal of Language and Linguistics Large-Scale Dictionary Construction for Foreign Language Tutoring and Interlingual Machine Translation

ﺑﺮﺍی ﻣﻄﺎﻟﻌﻪ ﺑیﺸﺘﺮ [1] Chris Quirk, ” Training a Sentence-Level Machine Translation Confidence Measure”, May 2004. [2] Einat Minkov, Kristina Toutanova, Hisami Suzuki Generating , ”Complex Morphology for Machine Translation”, June 2007. [3] Kristina Toutanova , Hisami Suzuki, ” Generating Case Markers in Machine Translation”, April 2007. [4] Robert C. Moore, Chris Quirk , ” Faster Beam. Search Decoding for Phrasal Statistical Machine Translation. ” September 2007.

ﺑﺎ ﺗﺸکﺮ