Extracting references from political speech auto-transcripts

This paper presents an unsupervised method for counting references in noisy auto-transcribed political speeches. Transcriptions are vectorized using learned embeddings which are then clustered using