Cross-Lingual Speaker Identification Using Distant Supervision
This work addresses speaker identification for literary analysis, offering cross-lingual generalization but is incremental as it builds on existing language models with new supervision methods.
The paper tackled the problem of speaker identification in literary text by proposing a framework that uses distant supervision to train a cross-lingual language model, resulting in up to 9% accuracy improvement on English benchmarks and up to 4.7% on Chinese datasets.
Speaker identification, determining which character said each utterance in literary text, benefits many downstream tasks. Most existing approaches use expert-defined rules or rule-based features to directly approach this task, but these approaches come with significant drawbacks, such as lack of contextual reasoning and poor cross-lingual generalization. In this work, we propose a speaker identification framework that addresses these issues. We first extract large-scale distant supervision signals in English via general-purpose tools and heuristics, and then apply these weakly-labeled instances with a focus on encouraging contextual reasoning to train a cross-lingual language model. We show that the resulting model outperforms previous state-of-the-art methods on two English speaker identification benchmarks by up to 9% in accuracy and 5% with only distant supervision, as well as two Chinese speaker identification datasets by up to 4.7%.