Date of Defense

Spring 4-15-2011

Department

Foreign Languages (to 2012)

First Advisor

Nicholas A. Andreadis, Lee Honors College

Second Advisor

Xiaojun Wang Foreign Languages

Third Advisor

Thomas Kostrzewa, Haenicke Institute for Global Education

Abstract

Higher demand for convenient translation methods dictates the necessity for improved machine translation programs. This paper details the theoretical approach and requirements for retrofitting a chess engine, including its algorithms and structural base, into a functional translation engine. The reason this approach works is that both chess engines and translation engines share similar programming roots despite focusing on two wildly different sections of computer programming and A.I. studies. Adaptation of the chess engine for machine translation would not prove too difficult given the inherent similarities in the two processes. Using statistical translation to keep track of a sentence's "score" will allow the program to return multiple, high-rated scores, which allows both efficient reader understanding and the potential for the program to educate itself more efficiently. The search tree can simply be modified to create sentence structures as opposed to chess board positions. The databases will need to be extended enormously to accept idioms and dictionaries, with search tree trimming functions redone to meet the new requirements of the system.

Comments

Scott Friesner, Lee Honors College

Access Setting

Honors Thesis-Campus Only

Restricted to Campus until

10-2031

Share

COinS