宁波东方理工大学自然语言处理课题组

我们是宁波东方理工大学自然语言处理课题组(EIT-NLP)。依托的宁波东方理工大学,坐落于孕育过王阳明、黄宗羲等思想大家与屠呦呦等科学大师的宁波 —— 这座东南沿海的港口名城与历史文化之城。学校以 “服务国家发展、推动社会进步” 为使命,是一所社会力量举办、省市共同建设、国家重点支持的高起点、小而精、创新型、国际化的新型研究型大学。

新闻

MultiConIR: Towards Multi-Condition Information Retrieval
Xuan Lu, Sifan Liu, Bochao Yin, Yongqi Li, Xinghao Chen, Hui Su, Yaohui Jin, Wenjun Zeng, Xiaoyu Shen
EMNLP 2025 Findings  ·  04 Sep 2025  ·  arxiv:2503.08046v3
Multimodal Language Models See Better When They Look Shallower
Haoran Chen, Junyan Lin, Xinhao Chen, Yue Fan, Xin Jin, Hui Su, Jianfeng Dong, Jinlan Fu, Xiaoyu Shen
EMNLP 2025  ·  30 Apr 2025  ·  arxiv:2504.21447
VisiPruner: Decoding Discontinuous Cross-Modal Dynamics for Efficient Multimodal LLMs
Yingqi Fan, Anhao Zhao, Jinlan Fu, Junlong Tong, Hui Su, Yijie Pan, Wei Zhang, Xiaoyu Shen
EMNLP 2025  ·  04 Sep 2025  ·  /
PricingLogic: Evaluating LLMs Reasoning on Complex Tourism Pricing Tasks
Yunuo Liu, Zena Al Khalili, Dawei Zhu, Yanjun Chen, Dietriech Klakow, Xiaoyu Shen
EMNLP 2025  ·  04 Sep 2025  ·  /
Auto-SLURP: A Benchmark Dataset for Evaluating Multi-Agent Frameworks in Smart Personal Assistant
Lei Shen, Xiaoyu Shen
EMNLP 2025 Findings  ·  25 Apr 2025  ·  arxiv:2504.18373
Context Guided Transformer Entropy Modeling for Video Compression
Junlong Tong, Wei Zhang, Yaohui Jin, Xiaoyu Shen
ICCV 2025  ·  03 Aug 2025  ·  arxiv:2508.01852
LLM as Effective Streaming Processor: Bridging Streaming-Batch Mismatches with Group Position Encoding
Junlong Tong, Jinlan Fu, Zixuan Lin, Yingqi Fan, Anhao Zhao, Hui Su, Xiaoyu Shen
EMNLP 2025 Findings  ·  29 May 2025  ·  arxiv:2505.16983
SkipGPT: Each Token is One of a Kind
Anhao Zhao, Fanghua Ye, Yingqi Fan, Junlong Tong, Zhiwei Fei, Hui Su, Xiaoyu Shen
ICML 2025  ·  04 Jan 2025  ·  arxiv:2506.04179
Scaling Under-Resourced TTS: A Data-Optimized Framework with Advanced Acoustic Modeling for Thai
Yizhong Geng, Jizhuo Xu, Zeyu Liang, Jinghan Yang, Xiaoyi Shi, Xiaoyu Shen
ACL 2025 (Industry Track)  ·  10 Apr 2025  ·  arxiv:2504.07858
Unveiling the Key Factors for Distilling Chain-of-Thought Reasoning
Xinghao Chen, Zhijing Sun, Wenjin Guo, Miaoran Zhang, Yanjun Chen, …, Hui Su, Yijie Pan, Dietrich Klakow, Wenjie Li, Xiaoyu Shen
ACL 2025 Findings  ·  27 May 2025  ·  arxiv:2502.18001
Multi-Layer Visual Feature Fusion in Multimodal LLMs: Methods, Analysis, and Best Practices
Junyan Lin, Haoran Chen, Yue Fan, Yingqi Fan, Xin Jin, Hui Su, Jinlan Fu, Xiaoyu Shen
CVPR 2025  ·  08 Mar 2025  ·  arxiv:2503.06063
Beyond Content Relevance: Evaluating Instruction Following in Retrieval Models
Jianqun Zhou, Yuanlei Zheng, Wei Chen, Qianqian Zheng, Hui Su, Wei Zhang, Rui Meng, Xiaoyu Shen
ICLR 2025  ·  05 Mar 2025  ·  arxiv:2410.23841

About us

We are the Natural Language Processing Group (NLP), Eastern institute of technology (EIT), Ningbo. Ningbo is a major port city and historical cultural center on China’s southeastern coast. The institute is committed to gathering top global talent and aims to establish itself as a high-level, innovative, and international research university. Its goal is to become a world-class academic institution in the short term and lead the development of key strategic disciplines in the country. The institute promotes an academic environment of equality, openness, and freedom, with the ambition of becoming a nonprofit, cutting-edge research institution driving transformative research.

At the NLP group, we align with EIT’s mission to drive transformative research. Our work focuses on advancing natural language processing through innovative approaches that enable machines to deeply understand, generate, and reason with human language. We tackle diverse areas such as information retrieval, multi-modal learning, and chain-of-thought reasoning. A core objective is to develop highly efficient algorithms that deliver state-of-the-art performance while minimizing computational and resource demands. We prioritize reducing dependence on large-scale human annotation, speeding up training processes, and cutting down inference costs, ultimately aiming to create robust, resource-efficient models that are accessible and effective across a wide range of applications.

For more updates, follow us on our official WeChat public account: EIT-NLP.

EIT-NLP 微信公众号二维码

Github

Our lab focus on open science and will release our code, model and dataset at VISIT OUR GITHUB