Presentation is loading. Please wait.

Presentation is loading. Please wait.

Syntactic Category Prediction for Improving Translation Quality in English-Korean Machine Translation Sung-Dong Kim, Dept. of Computer Engineering, Hansung.

Similar presentations


Presentation on theme: "Syntactic Category Prediction for Improving Translation Quality in English-Korean Machine Translation Sung-Dong Kim, Dept. of Computer Engineering, Hansung."— Presentation transcript:

1 Syntactic Category Prediction for Improving Translation Quality in English-Korean Machine Translation Sung-Dong Kim, Dept. of Computer Engineering, Hansung University, Republic of Korea

2 Abstract Syntactic category prediction In this work
Guide parser to identify segment relationships Select correct parsing results for each segment In this work Feature design Decision tree generation  generation of prediction rules Prediction accuracy comparison Contribution to improving translation quality Syntactic Category Prediction

3 1. Introduction Long sentence translation Bad translation results
comma-separated phrases/clauses Sentence segmentation & segment parsing identify segment relationships selection of correct segment parsing results Syntactic Category Prediction

4 Syntactic category prediction of each segment guide parsing
get better translation results Syntactic Category Prediction

5 2. Related Works & Parsing Method
A small TCL interpreter, which can be linked into the code, interprets the strings 작은(small) TCL 인터프리터(interpreter), 그 코드(code)로 연결(link)될 수 있는 것(which), 그 문자열(strings)을 해석한다(interprets). 그 코드(code)로 연결(link)될 수 있는 작은(small) TCL 인터프리터(interpreter)는 그 문자열(strings)을 해석한다(interprets). Syntactic Category Prediction

6 Parsing Steps Syntactic Category Prediction

7 3. Generation of Prediction Rules/Functions
Syntactic Category Prediction

8 3. Feature, Training Data First word First tag Second word Second tag
Last tag Segment length Clue for clause Pierre/NNP Vinken/NNP NP 61/CD years/NNS old/JJ AJP will/MD join/VB the/DT board/NN as/IN a/DT nonexecutive/JJ director/NN Nov./NNP 29/CD VP 2, NNP, NNP, *NONE*, 2, NNP, 2, NP 0, CD, JJ, *NONE*, 4139, NNS, 3, AJP 4085, MD, CD, MD, 1920, VB, 10, VP Syntactic Category Prediction

9 3.Prediction Rules if (!strcmp(szFirstWord,“that”) && check_IN(firstTrees)) { setTargetCategory( THAT_CLAUSE ); return; } if (!strcmp(szFirstWord,"although") && check_IN(firstTrees)) { setTargetCategory( SUBCL ); return; } if (check_IN(firstTrees) && checkClauseExist(_NONE) && check_NNP(secondTrees)) { setTargetCategory( PP ); return; } if (check_CD(firstTrees) && check_NONE(lastTrees)) { setTargetCategory( NP ); return; } Syntactic Category Prediction

10 4. Experiments Syntactic Category Prediction


Download ppt "Syntactic Category Prediction for Improving Translation Quality in English-Korean Machine Translation Sung-Dong Kim, Dept. of Computer Engineering, Hansung."

Similar presentations


Ads by Google