Logits-Constrained Framework with RoBERTa for Ancient Chinese NER

TL;DR

This paper introduces a Logits-Constrained framework using RoBERTa for Ancient Chinese NER, enhancing label transition validity and outperforming traditional methods on the EvaHan 2025 benchmark.

Contribution

It proposes a novel two-stage LC framework with a differentiable decoder and a model selection criterion tailored for Ancient Chinese NER tasks.

Findings

LC outperforms CRF and BiLSTM models

Effective in high-label and large-data scenarios

Provides practical model selection guidance

Abstract

This paper presents a Logits-Constrained (LC) framework for Ancient Chinese Named Entity Recognition (NER), evaluated on the EvaHan 2025 benchmark. Our two-stage model integrates GujiRoBERTa for contextual encoding and a differentiable decoding mechanism to enforce valid BMES label transitions. Experiments demonstrate that LC improves performance over traditional CRF and BiLSTM-based approaches, especially in high-label or large-data settings. We also propose a model selection criterion balancing label complexity and dataset size, providing practical guidance for real-world Ancient Chinese NLP tasks.