File model approach to optimize the performance of Tree Adjoining Grammar based Machine Translation

Abstract

The growing pace of information technology demands fast operation for communication and other related applications. After having a successful Machine Translation System [MTS], it has been felt to optimize the performance of Machine Translation for its real-time uses. The considered MTS is Tree Adjoining Grammar [TAG] based system. An approach has been experimented to use File model instead of Database model for fast streaming of grammar into memory and operation. This model provides an efficient and a systematic way of encapsulating language resource with engineering solution to develop the speedy MTS. The computational experiments demonstrate that substantial performance in terms of time and memory has been obtained by using this approach.

Unveiling the Power of TAG Using Statistical Parsing for Natural Languages

“…We have also proposed Virtual research Lab [6] for TAG related research where we have described extension of TAG Derivation. Earlier, we had also made effort to improve Performance of TAG based Machine Translation [7]. We found one close work related to our research by Vijay-Shanker [8] where they have described LTAGs and their application using statistical parsing.…”

“…We found that the computational complexity of TAG is extremely high because of structural ambiguity in natural languages and differences between the elementary trees of various languages. Our earlier research also described complexity and challenges beneath parsing-generation process [2]. To address the challenges in the existing TAG Parser, we explored an alternative approach employing a probabilistic model using advanced algorithms, and we observed substantial enhancements in both the performance and efficiency of numerous natural language processing (NLP) applications Therefore, we explored Statistical Parsing algorithm to facilitate the translation of sentences from one language to another.…”