Political corpus creation through automatic speech recognition on EU debates

TL;DR

This paper develops an accurate automatic speech recognition pipeline for EU parliamentary debates, creating a large transcribed corpus to facilitate political research and analysis.

Contribution

It introduces a domain-specific ASR model using unsupervised adaptation of Wav2vec2.0, significantly improving transcription accuracy for EU debate recordings.

Findings

Domain-specific acoustic and language models reduce WER from 28.22% to 17.95%.

Adding domain-specific terms did not improve ASR performance.

The corpus enables effective topic modeling for political analysis.

Abstract

In this paper, we present a transcribed corpus of the LIBE committee of the EU parliament, totalling 3.6 Million running words. The meetings of parliamentary committees of the EU are a potentially valuable source of information for political scientists but the data is not readily available because only disclosed as speech recordings together with limited metadata. The meetings are in English, partly spoken by non-native speakers, and partly spoken by interpreters. We investigated the most appropriate Automatic Speech Recognition (ASR) model to create an accurate text transcription of the audio recordings of the meetings in order to make their content available for research and analysis. We focused on the unsupervised domain adaptation of the ASR pipeline. Building on the transformer-based Wav2vec2.0 model, we experimented with multiple acoustic models, language models and the addition of domain-specific terms. We found that a domain-specific acoustic model and a domain-specific language model give substantial improvements to the ASR output, reducing the word error rate (WER) from 28.22 to 17.95. The use of domain-specific terms in the decoding stage did not have a positive effect on the quality of the ASR in terms of WER. Initial topic modelling results indicated that the corpus is useful for downstream analysis tasks. We release the resulting corpus and our analysis pipeline for future research.