Cisco at AAAI-CAD21 shared task: Predicting Emphasis in Presentation Slides using Contextualized Embeddings

TL;DR

This paper presents two approaches using contextualized embeddings to predict emphasis in presentation slides, achieving top leaderboard rankings in a shared task.

Contribution

The paper introduces a dual approach employing BiLSTM-ELMo and transformer models for emphasis prediction in slides, advancing methods in slide content analysis.

Findings

Achieved 0.518 score, ranking 3rd during evaluation.

Achieved 0.543 score, ranking 1st post-evaluation.

Demonstrated effectiveness of transformer-based models in emphasis prediction.

Abstract

This paper describes our proposed system for the AAAI-CAD21 shared task: Predicting Emphasis in Presentation Slides. In this specific task, given the contents of a slide we are asked to predict the degree of emphasis to be laid on each word in the slide. We propose 2 approaches to this problem including a BiLSTM-ELMo approach and a transformers based approach based on RoBERTa and XLNet architectures. We achieve a score of 0.518 on the evaluation leaderboard which ranks us 3rd and 0.543 on the post-evaluation leaderboard which ranks us 1st at the time of writing the paper.