# Exploring sentence informativeness

## TL;DR

This paper investigates the concept of sentence informativeness and its impact on improving word embeddings, proposing classifiers to predict informativeness and demonstrating their influence on embedding quality.

## Contribution

It introduces sentence-level classifiers for informativeness prediction and explores their use in enhancing word representation quality from limited data.

## Key findings

- Classifiers can predict sentence informativeness effectively.
- Informativeness measures differ from each other.
- Using classifiers' predictions improves word embedding quality.

## Abstract

This study is a preliminary exploration of the concept of informativeness -how much information a sentence gives about a word it contains- and its potential benefits to building quality word representations from scarce data. We propose several sentence-level classifiers to predict informativeness, and we perform a manual annotation on a set of sentences. We conclude that these two measures correspond to different notions of informativeness. However, our experiments show that using the classifiers' predictions to train word embeddings has an impact on embedding quality.