# IIT (BHU) Varanasi at MSR-SRST 2018: A Language Model Based Approach for Natural Language Generation

## TL;DR

This paper presents a system for surface realization in natural language generation that combines LSTM-based word reinflection with language model techniques for word order prediction, improving sentence formation from structured data.

## Contribution

The paper introduces a hybrid approach using LSTM and language models for surface realization, addressing word reinflection and word order prediction in NLG tasks.

## Key findings

- Effective combination of LSTM and language models improves sentence generation.
- Two sub-approaches in language modeling enhance the accuracy of surface realization.
- System performs well on the SRST'18 shared task dataset.

## Abstract

This paper describes our submission system for the Shallow Track of Surface Realization Shared Task 2018 (SRST'18). The task was to convert genuine UD structures, from which word order information had been removed and the tokens had been lemmatized, into their correct sentential form. We divide the problem statement into two parts, word reinflection and correct word order prediction. For the first sub-problem, we use a Long Short Term Memory based Encoder-Decoder approach. For the second sub-problem, we present a Language Model (LM) based approach. We apply two different sub-approaches in the LM Based approach and the combined result of these two approaches is considered as the final output of the system.