Papers by Shuaijie She
A Data-Efficient Path to Multilingual LLMs: Language Expansion via Post-training PARAM𝛥 Integration into Upcycled MoE (2026.acl-long)
Copied to clipboard
Hao Zhou, Tianhao Li, Zhijun Wang, Shuaijie She, Linjuan Wu, Hao-Ran Wei, Baosong Yang, Jiajun Chen, Shujian Huang
| Challenge: | Large Language Models (LLMs) are expensive and require extensive Continued Pre-Training and data-intensive alignment to expand. |
| Approach: | They propose a method which upcycles a dense model into a Mixture-of-Experts architecture, allocating different experts to different languages. |
| Outcome: | Experiments show that the proposed model upcycles a dense model into a Mixture-of-Experts(MoE) architecture, allocating different experts to different languages. |
MAPO: Advancing Multilingual Reasoning through Multilingual-Alignment-as-Preference Optimization (2024.acl-long)
Copied to clipboard
| Challenge: | Existing models exhibit inconsistent reasoning abilities across different languages . existing models lack consistency across languages due to imbalance of training data . |
| Approach: | They propose a multilingual alignment-as-preference optimization framework to align reasoning processes in other languages with the dominant language. |
| Outcome: | The proposed framework improves multilingual reasoning across languages on three benchmarks. |
Improved Pseudo Data for Machine Translation Quality Estimation with Constrained Beam Search (2023.emnlp-main)
Copied to clipboard
Xiang Geng, Yu Zhang, Zhejian Lai, Shuaijie She, Wei Zou, Shimin Tao, Hao Yang, Jiajun Chen, Shujian Huang
| Challenge: | evaluating the quality of machine translation outputs becomes increasingly essential with the rapid development of machine language (MT). |
| Approach: | They propose to generate pseudo data using the MT model with constrained beam search (CBSQE) they propose to preserve the reference parts with high MT probabilities as correct translations . |
| Outcome: | The proposed model outperforms strong baselines in both supervised and unsupervised settings. |
Multilingual Contrastive Decoding via Language-Agnostic Layers Skipping (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Decoding by contrasting layers (DoLa) is designed to improve the generation quality of large language models (LLMs) however, this approach does not work well on non-English tasks. |
| Approach: | They propose a contrastive decoding algorithm that uses amateur logits to contrast with the output of an expert model's early exit logits. |
| Outcome: | The proposed method outperforms baselines and significantly improves chain-of-thought reasoning accuracy across 11 languages. |
R-PRM: Reasoning-Driven Process Reward Modeling (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing Process Reward Models (PRMs) output evaluation scores directly, limiting both learning efficiency and evaluation accuracy. |
| Approach: | They propose a Reasoning-Driven Process Reward Modeling (R-PRM) which activates inherent reasoning to enhance process-level evaluation. |
| Outcome: | The proposed model outperforms baseline models on ProcessBench and PRMBench by 13.9 and 8.5 F1 scores. |
Question Translation Training for Better Multilingual Reasoning (2024.findings-acl)
Copied to clipboard
| Challenge: | Large language models have shown compelling performance on reasoning tasks but they tend to perform much worse in languages other than English. |
| Approach: | They propose to train a model to translate reasoning questions into English by fine tuning on X-English parallel question data. |
| Outcome: | The proposed approach improves on LLaMA2-13B on the MGSM and MSVAMP multilingual reasoning benchmarks. |
Exploring the Factual Consistency in Dialogue Comprehension of Large Language Models (2024.naacl-long)
Copied to clipboard
| Challenge: | LLMs generate responses following user's instructions, which requires high dialogue comprehension ability. |
| Approach: | They propose to evaluate LLMs' dialogue comprehension ability using a dialogue summarization task to derive factual questions from the generated summaries and use them as a more flexible measurement of dialogue comprehension. |
| Outcome: | The proposed model reduces the error rate by 11% on the dialogue summarization task. |
Improving Long-Context Translation via Self-Supervised Dual Learning (2026.acl-long)
Copied to clipboard
| Challenge: | Large language models with long context windows suffer from catastrophic information distortion, undermining the strict faithfulness required for translation. |
| Approach: | They propose a self-supervised post-training framework that improves long-document translation reliability via round-trip consistency. |
| Outcome: | The proposed framework improves long-document translation reliability via round-trip consistency. |