Decoding Text Spans for Efficient and Accurate Named-Entity Recognition
This work addresses efficiency bottlenecks in NER for industrial and on-device deployments, offering an incremental improvement over existing span-based methods.
The paper tackled the problem of high inference cost in span-based Named Entity Recognition (NER) by proposing SpanDec, which reduces redundant computation and prunes unlikely candidates, achieving competitive accuracy while improving throughput and reducing computational cost for high-volume and on-device applications.
Named Entity Recognition (NER) is a key component in industrial information extraction pipelines, where systems must satisfy strict latency and throughput constraints in addition to strong accuracy. State-of-the-art NER accuracy is often achieved by span-based frameworks, which construct span representations from token encodings and classify candidate spans. However, many span-based methods enumerate large numbers of candidates and process each candidate with marker-augmented inputs, substantially increasing inference cost and limiting scalability in large-scale deployments. In this work, we propose SpanDec, an efficient span-based NER framework that targets this bottleneck. Our main insight is that span representation interactions can be computed effectively at the final transformer stage, avoiding redundant computation in earlier layers via a lightweight decoder dedicated to span representations. We further introduce a span filtering mechanism during enumeration to prune unlikely candidates before expensive processing. Across multiple benchmarks, SpanDec matches competitive span-based baselines while improving throughput and reducing computational cost, yielding a better accuracy-efficiency trade-off suitable for high-volume serving and on-device applications.