Topic Model Robustness to Automatic Speech Recognition Errors in Podcast Transcripts

Abstract

For a multilingual podcast streaming service, it is critical to be able to deliver relevant content to all users independent of language. Podcast content relevance is conventionally determined using various metadata sources. However, with the increasing quality of speech recognition in many languages, utilizing automatic transcriptions to provide better content recommendations becomes possible. In this work, we explore the robustness of a Latent Dirichlet Allocation topic model when applied to transcripts created by an automatic speech recognition engine. Specifically, we explore how increasing transcription noise influences topics obtained from transcriptions in Danish; a low resource language. First, we observe a baseline of cosine similarity scores between topic embeddings from automatic transcriptions and the descriptions of the podcasts written by the podcast creators. We then observe how the cosine similarities decrease as transcription noise increases and conclude that even when automatic speech recognition transcripts are erroneous, it is still possible to obtain high-quality topic embeddings from the transcriptions.

References in corpus (11)

• SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition
• Probabilistic Latent Semantic Analysis
• Libri-Light: A Benchmark for ASR with Limited or No Supervision
• Synthetic and Natural Noise Both Break Neural Machine Translation
• Improved Noisy Student Training for Automatic Speech Recognition
• Common Voice: A Massively-Multilingual Speech Corpus
• Pushing the Limits of Semi-Supervised Learning for Automatic Speech Recognition
• First-Pass Large Vocabulary Continuous Speech Recognition using Bi-Directional Recurrent DNNs
• How Many Topics? Stability Analysis for Topic Models
• An Approach to Improve Robustness of NLP Systems against ASR Errors
• Current Challenges and Future Directions in Podcast Information Access