BERTScore: Evaluating Text Generation with BERT Article Swipe

PDF

Related Concepts

security token paraphrase computer science machine translation similarity (geometry) artificial intelligence sentence metric (unit) selection (genetic algorithm) natural language processing task (project management) closed captioning machine learning speech recognition image (mathematics) economics operations management computer security management

Tianyi Zhang , Varsha Kishore , Felix Wu , Kilian Q. Weinberger , Yoav Artzi ·

YOU? · · 2019 · Open Access · · DOI: https://doi.org/10.48550/arxiv.1904.09675 · OA: W2936695845

We propose BERTScore, an automatic evaluation metric for text generation. Analogously to common metrics, BERTScore computes a similarity score for each token in the candidate sentence with each token in the reference sentence. However, instead of exact matches, we compute token similarity using contextual embeddings. We evaluate using the outputs of 363 machine translation and image captioning systems. BERTScore correlates better with human judgments and provides stronger model selection performance than existing metrics. Finally, we use an adversarial paraphrase detection task to show that BERTScore is more robust to challenging examples when compared to existing metrics.