Research map / Large language models / Language, safety and interpretability
From Good to Best: Two-Stage Training for Cross-Lingual Machine Reading Comprehension
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use two-stage hard-learning and answer-aware contrastive learning | XLM-Rbase · MLQA zero-shot (Avg.) | F1: 64.14 -> 66.00 | Helped |
| Use two-stage hard-learning and answer-aware contrastive learning | XLM-Rbase · MLQA zero-shot (Avg.) | EM: 46.00 -> 48.88 | Helped |
| Use two-stage hard-learning and answer-aware contrastive learning | XLM-Rbase · MLQA translate-train (Avg.) | F1: 65.45 -> 67.16 | Helped |
| Use two-stage hard-learning and answer-aware contrastive learning | XLM-Rbase · MLQA translate-train (Avg.) | EM: 48.25 -> 50.44 | Helped |
| Use two-stage hard-learning and answer-aware contrastive learning | XLM-Rbase · XQUAD zero-shot (Avg.) | F1: +1.52% | Helped |
| Use two-stage hard-learning and answer-aware contrastive learning | XLM-Rbase · XQUAD zero-shot (Avg.) | EM: +2.32% | Helped |
| Use two-stage hard-learning and answer-aware contrastive learning | XLM-Rbase · XQUAD translate-train (Avg.) | F1: 75.03 -> 76.66 | Helped |
| Use two-stage hard-learning and answer-aware contrastive learning | XLM-Rbase · XQUAD translate-train (Avg.) | EM: 58.96 -> 60.66 | Helped |
| Add hard-learning (HL) (ablation) | Full method · MLQA translate-train, Spanish | F1: 67.64 -> 69.04 | Helped |
| Add hard-learning (HL) (ablation) | Full method · MLQA translate-train, Arabic | F1: 57.10 -> 58.54 | Helped |
| Add hard-learning (HL) (ablation) | Full method · MLQA translate-train, Vietnamese | F1: 66.00 -> 67.92 | Helped |
| Add answer-aware contrastive learning (AA-CL) (ablation) | Full method · MLQA translate-train, Spanish | F1: 67.70 -> 69.04 | Helped |
Closest papers by meaning
- A Graph Fusion Approach for Cross-Lingual Machine Reading Comprehension (AAAI 2023 · similarity 0.91)
- Zero-Shot Cross-Lingual Machine Reading Comprehension via Inter-sentence Dependency Graph (AAAI 2022 · similarity 0.89)
- Assessing the Benchmarking Capacity of Machine Reading Comprehension Datasets (AAAI 2020 · similarity 0.88)
- XCOT: Cross-lingual Instruction Tuning for Cross-lingual Chain-of-Thought Reasoning (AAAI 2025 · similarity 0.87)
- Linguini: A benchmark for language-agnostic linguistic reasoning (NeurIPS 2025 · similarity 0.86)
- MCIF: Multimodal Crosslingual Instruction-Following Benchmark from Scientific Talks (ICLR 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.