Papers by Kai Shi
UAQFact: Evaluating Factual Knowledge Utilization of LLMs on Unanswerable Questions (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing datasets to assess LLMs' performance on unanswerable questions lack factual knowledge support. |
| Approach: | They propose a bilingual unanswerable question dataset with auxiliary factual knowledge created from a Knowledge Graph and two new tasks to measure LLMs' ability to utilize internal and external factual information. |
| Outcome: | The proposed datasets show that LLMs do not consistently perform well even when they have factual knowledge stored. |
wav2vec-S: Adapting Pre-trained Speech Models for Streaming (2024.findings-acl)
Copied to clipboard
| Challenge: | Pre-trained speech models have advanced speech-related tasks, including speech recognition and translation. |
| Approach: | They propose a pre-trained speech model that incorporates modifications to ensure consistent speech representations during training and inference phases for streaming speech inputs. |
| Outcome: | The proposed model outperforms baseline models on speech recognition and translation tasks and achieves a superior balance between quality and latency. |
LLMs Can Achieve High-quality Simultaneous Machine Translation as Efficiently as Offline (2025.findings-acl)
Copied to clipboard
| Challenge: | Large language models perform well in offline machine translation when the complete source sentence is provided . however, in many real scenarios, the source tokens arrive in a streaming manner and simultaneous machine translation is required . |
| Approach: | They propose a new paradigm that includes constructing supervised fine-tuning data for simultaneous machine translation (SiMT) to achieve SiMT, source and target tokens are rearranged into interleaved sequences, separated by special tokens according to varying latency requirements. |
| Outcome: | The proposed approach achieves state-of-the-art performance across various SiMT benchmarks and evaluation metrics while maintaining efficient auto-regressive decoding. |
TongGu: Mastering Classical Chinese Understanding with Knowledge-Grounded Large Language Models (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have demonstrated remarkable capability in Natural Language Processing (NLP), but struggle with Classical Chinese Understanding (CCU) Existing models, including general-purpose and preliminary LLMs, lack the ability to address CCU in data-demanding and knowledge-intensive tasks. |
| Approach: | They propose to use a classical Chinese corpora-based instruction-tuning dataset to unlock the full CCU potential of LLMs. |
| Outcome: | The proposed model unlocks the full CCU potential of LLMs by preserving its foundational knowledge while maintaining redundancy-aware tuning (RAT) and CCU-RAG. |
Computer Assisted Translation with Neural Quality Estimation and Automatic Post-Editing (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Using neural machine translation to approximate human parity is difficult due to the lack of parallel training corpora. |
| Approach: | They propose an end-to-end deep learning framework for quality estimation and automatic post-editing of machine translation output. |
| Outcome: | The proposed framework achieves state-of-the-art performance on the English–German dataset and human translators can significantly expedite their post-editing processing with the model. |
Memory-Guided Hard Data Augmentation for Multimodal Named Entity Recognition (2026.findings-acl)
Copied to clipboard
Xinyu Liu, Kai fu, Yinghan Shi, Quanyou Chu, Ming Du, Hongya Wang, Xiaojun Meng, Jiansheng Wei, Yanghua Xiao, Bo Xu
| Challenge: | Existing methods for Named Entity Recognition (NER) ignore the internal state of the target model. |
| Approach: | They propose a framework to repair model-specific errors by using a model-based approach . they employ cross-validation to identify model- specific Hard Data and a memory tree to induce macro-level error patterns from micro-level failures. |
| Outcome: | The proposed framework yields significant performance gains on Twitter and other platforms. |
Review-Instruct: A Review-Driven Multi-Turn Conversations Generation Method for Large Language Models (2025.findings-acl)
Copied to clipboard
Jiangxu Wu, Cong Wang, Tianhuang Su, Lin Haozhi, JunYang JunYang, Zhangchao Zhangchao, Binqiang Pan, SongpanYang SongpanYang, Mingpeng Mingpeng, Kai Shi, Zixian Li
| Challenge: | Existing methods for generating multi-turn dialogue data struggle to ensure both diversity and quality in instructions. |
| Approach: | They propose a framework that synthesizes multi-turn conversations through an iterative "Ask-Respond-Review" process involving three agent roles: a Candidate, multiple Reviewers, and a Chairman. |
| Outcome: | The proposed framework synthesizes multi-turn conversations through an iterative "Ask-Respond-Review" process involving three agent roles: a Candidate, multiple Reviewers, and a Chairman. |
Mixture of Heterogeneous Grouped Experts for Language Modeling (2026.acl-industry)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) based on Mixture-of-Experts (MoE) enforce uniform expert sizes, creating a rigidity that fails to align computational costs with varying token-level complexity. |
| Approach: | They propose a mixture of heterogeneous grouped experts (MoHGE) that allows for flexible, resource-aware expert combinations. |
| Outcome: | The proposed model matches the performance of existing Mixture-of-Experts architectures while maintaining balanced GPU utilization. |
STEER-BENCH: A Benchmark for Evaluating the Steerability of Large Language Models (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large language models can adapt outputs to align with community-specific norms, perspectives and communication styles. |
| Approach: | They propose a benchmark to assess community-specific steering using contrasting reddit communities. |
| Outcome: | STEER-BENCH assesses how well large language models understand community-specific instructions, their resilience to adversarial steering attempts, and their ability to accurately represent cultural and ideological perspectives. |
CLUE: A Chinese Language Understanding Evaluation Benchmark (2020.coling-main)
Copied to clipboard
Liang Xu, Hai Hu, Xuanwei Zhang, Lu Li, Chenjie Cao, Yudong Li, Yechen Xu, Kai Sun, Dian Yu, Cong Yu, Yin Tian, Qianqian Dong, Weitang Liu, Bo Shi, Yiming Cui, Junyi Li, Jun Zeng, Rongzhao Wang, Weijian Xie, Yanting Li, Yina Patterson, Zuoyu Tian, Yiwen Zhang, He Zhou, Shaoweihua Liu, Zhe Zhao, Qipeng Zhao, Cong Yue, Xinrui Zhang, Zhengliang Yang, Kyle Richardson, Zhenzhong Lan
| Challenge: | Existing language evaluation benchmarks for English are limited to English . lack of such benchmarks makes it difficult to replicate success in other languages . |
| Approach: | They introduce a large-scale Chinese language understanding evaluation benchmark . the benchmark uses a set of current state-of-the-art pre-trained Chinese models . |
| Outcome: | The first large-scale Chinese Language Understanding Evaluation (CLUE) benchmark is released . the benchmark evaluates models across a wide range of tasks on original Chinese text . existing language evaluation benchmarks are mostly limited to English . |
Reinforcement Learning on Pre-Training Data (2026.acl-long)
Copied to clipboard
Siheng Li, Kejiao Li, Zenan Xu, Guanhua Huang, Kun Li, Haoyuan Wu, null Wujiajia, Zihao Zheng, Chenchen Zhang, Kun Shi, Xue Gong, Qi Yi, Ruibin Xiong, Tingqiang Xu, Yuhao Jiang, Jianfeng Yan, Yuyuan Zeng, Guanghui Xu, Jinbao Xue, Zhijiang xu, Zheng Fang, Shuai LI, Qibin Liu, Xiaoxue Li, Zhuoyu Li, Yangyu Tao, Fei Gao, Cheng Jiang, Bochao Wang, Kai Liu, Jianchen Zhu, Wai Lam, Bo Zhou, Di Wang
| Challenge: | Recent progress in large language models is driven by scaling of training compute through pre-training with nexttoken prediction (NTP) or post-training (RL) Pre-training using NTP enables models to acquire extensive knowledge and skills from general data, but it suffers from data inefficiency and catastrophic forgetting in continual learning settings. |
| Approach: | They propose to scale training compute through pre-training with next-token prediction (NTP) or post-training by scaling reinforcement learning (RL) to improve learning from general data. |
| Outcome: | Experiments on multiple benchmarks and models show that the proposed approach improves continual pre-training and provides a strong foundation for post-training on Qwen3-8B-Base. |
iAgent: LLM Agent as a Shield between User and Recommender Systems (2025.findings-acl)
Copied to clipboard
Wujiang Xu, Yunxiao Shi, Zujie Liang, Xuying Ning, Kai Mei, Kun Wang, Xi Zhu, Min Xu, Yongfeng Zhang
| Challenge: | Traditional recommender systems focus on the user-platform paradigm, where users are directly exposed under the control of the platform's recommendation algorithms. |
| Approach: | They propose a user-agent-platform paradigm where agent serves as the protective shield between user and recommender system that enables indirect exposure. |
| Outcome: | The proposed model improves 16.6% over baselines on four datasets and mitigates echo chamber effects and reduces model bias in disadvantaged users. |
How Susceptible are Large Language Models to Ideological Manipulation? (2024.emnlp-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have the potential to exert substantial influence on public perceptions and interactions with information. |
| Approach: | They examine how LLMs can learn and generalize ideological biases from their instruction-tuning data. |
| Outcome: | The LLMs show a startling ability to absorb ideology from one topic and generalize it to even unrelated ones. |
DMRetriever: A Family of Models for Improved Text Retrieval in Disaster Management (2026.acl-long)
Copied to clipboard
| Challenge: | Existing models fail to handle the varied search intents inherent to disaster management scenarios, resulting in inconsistent and unreliable performance. |
| Approach: | They propose a new series of dense retrieval models tailored for disaster management that train on a three-stage framework with unsupervised contrastive pre-training and difficulty-aware progressive instruction fine-tuning. |
| Outcome: | The proposed model outperforms baseline models by 13.3 times and 33 times over baselines with only 7.6% of their parameters. |
BSCodec: A Band-Split Neural Codec for High-Quality Universal Audio Reconstruction (2026.findings-eacl)
Copied to clipboard
| Challenge: | Neural audio codecs have enabled high-fidelity reconstruction of speech, music and sound . however, speech-optimized codec systems suffer degradation on music or sound if they ignore spectral differences . |
| Approach: | They propose a neural audio codec that splits the spectral dimension into separate bands and compresses each band independently. |
| Outcome: | Experimental results show that BSCodec achieves better reconstruction quality on music and sound compared to existing codecs. |
DaMo: Data Mixing Optimizer in Fine-tuning Multimodal LLMs for Mobile Phone Agents (2026.findings-acl)
Copied to clipboard
Kai Shi, Jun Yang, Ni Yang, Binqiang Pan, Qingsong Xie, null Zhangchao, Zhenyu Yang, Tianhuang Su, Haonan Lu
| Challenge: | Mobile Phone Agents (MPAs) have attracted huge attention due to their practicability in a multitude of scenarios. |
| Approach: | They propose a data mixture optimization solution that extrapolates optimal data mixtures from a trainable network. |
| Outcome: | The proposed model outperforms existing methods on open-source benchmarks and on open source benchmarks. |
Adapting Offline Speech Translation Models for Streaming with Future-Aware Distillation and Inference (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches to streaming speech translation use an offline model with a wait-k policy . however, there is a mismatch problem with an offline inference model trained with complete utterances . |
| Approach: | They propose an offline streaming speech translation model with wait-k policy to support different latency requirements. |
| Outcome: | The proposed model achieves better trade-offs between translation quality and latency than baselines. |
ReflectRM: Boosting Generative Reward Models via Self-Reflection within a Unified Judgment Framework (2026.acl-long)
Copied to clipboard
Kai Qin, Liangxin Liu, Yu Liang, Longzheng Wang, null Wangyan, Zhang Yueyang, Long Xia, Zhiyuan Sun, Houde Liu, Daiting Shi
| Challenge: | Existing methods for generating reward models focus on outcome-level supervision, neglecting analytical process quality, which constrains their potential. |
| Approach: | They propose a novel reward model that leverages self-reflection to assess analytical quality and enhance preference modeling. |
| Outcome: | The proposed model improves performance on four benchmarks and significantly mitigates positional bias. |
Natural Logic at the Core: Dynamic Rewards for Entailment Tree Generation (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to generating entailment trees lack logical consistency . static reward structures or intricate dependencies within multi-step reasoning are often ignored . |
| Approach: | They propose a method that integrates natural logic principles into reinforcement learning to guide entailment tree generation. |
| Outcome: | Experiments on EntailmentBank show that the proposed method improves interpretability and generalization. |
Prompting Large Language Models to Tackle the Full Software Development Lifecycle: A Case Study (2025.coling-main)
Copied to clipboard
Bowen Li, Wenhan Wu, Ziwei Tang, Lin Shi, John Yang, Jinyang Li, Shunyu Yao, Chen Qian, Binyuan Hui, Qicheng Zhang, Zhiyin Yu, He Du, Ping Yang, Dahua Lin, Chao Peng, Kai Chen
| Challenge: | Existing benchmarks focused on simplified or isolated aspects of coding, ignoring the full spectrum of programming challenges. |
| Approach: | They propose a case study that examines the performance of large language models across the entire software development lifecycle with four programming languages, multiple domains, and carefully designed and verified metrics for each task. |
| Outcome: | The proposed model performs across the entire software development lifecycle, including design, environment setup, implementation, acceptance testing, and unit testing. |
Safer-Instruct: Aligning Language Models with Automated Preference Data (2024.naacl-long)
Copied to clipboard
| Challenge: | annotating preference data by humans is resource-intensive and creativity-demanding . existing methods face limitations in data diversity and quality . |
| Approach: | They propose a pipeline for annotating large-scale preference data without human annotators. |
| Outcome: | The proposed pipeline outperforms models fine-tuned on human-annotated safety preference data while maintaining a competitive edge in downstream tasks. |
BrowseComp-Plus: A Fair and Disentangled Evaluation Benchmark for Deep Search Agents (2026.acl-long)
Copied to clipboard
Zijian Chen, Xueguang Ma, Shengyao Zhuang, Ping Nie, Kai Zou, Sahel Sharifymoghaddam, Andrew Liu, Joshua Green, Kshama Patel, Ruoxi Meng, Mingyi Su, Yanxi Li, Haoran Hong, Xinyu Shi, Xuye Liu, Hosna Oyarhoseini, Nandan Thakur, Crystina Zhang, Luyu Gao, Wenhu Chen, Jimmy Lin
| Challenge: | Existing benchmarks for deep search agents rely on blackbox web search APIs . dynamic and opaque web APIs hinder reproducibility and fair comparisons - authors . |
| Approach: | They propose a benchmark that employs a fixed corpus for controlled retrieval for deep search agents. |
| Outcome: | The new benchmark shows that agents that combine large language models with retrieval tools excel at complex, reasoning-intensive queries. |
Deciphering the Impact of Pretraining Data on Large Language Models through Machine Unlearning (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing studies have suggested that the composition of the pretraining corpus exerts a significant impact upon the performance of LLMs. |
| Approach: | They analyze the impact of 48 datasets from 5 major categories of pretraining data of Large Language Models and measure their impacts on LLMs using benchmarks about nine major categories. |
| Outcome: | The proposed analysis provides insights into the organization of data to support more efficient pretraining of Large Language Models. |
DAST: Difficulty-Adaptive Slow-Thinking for Large Reasoning Models (2025.emnlp-industry)
Copied to clipboard
Yi Shen, Jian Zhang, Jieyun Huang, Shuming Shi, Wenjing Zhang, Jiangze Yan, Ning Wang, Kai Wang, Zhaoxiang Liu, Shiguo Lian
| Challenge: | Recent advances in slow-thinking reasoning models have shown exceptional performance in complex reasoning tasks. |
| Approach: | They propose a framework that enables models to automatically adjust Chain-of-Thought (CoT) length based on problem difficulty. |
| Outcome: | The proposed framework penalizes inefficiency on simple problems while incentivizing deep reasoning for complex ones. |
D4: a Chinese Dialogue Dataset for Depression-Diagnosis-Oriented Chat (2022.emnlp-main)
Copied to clipboard
| Challenge: | Existing human-machine dialogue systems are not able to provide diagnostic information for depression diagnosis due to stigma associated with mental illness. |
| Approach: | They propose to construct a Chinese Dialogue Dataset for depression-diagnosis-oriented chat based on clinical depression diagnostic criteria. |
| Outcome: | The proposed system can be used to diagnose depression using a Chinese Dialogue Dataset. |
MCS-Bench: A Comprehensive Benchmark for Evaluating Multimodal Large Language Models in Chinese Classical Studies (2025.acl-long)
Copied to clipboard
| Challenge: | Multimodal Large Language Models (MLLMs) have advanced visual and language understanding, but their potential in Chinese Classical Studies (CCS) remains underexplored due to the lack of specialized benchmarks. |
| Approach: | They propose to develop a multimodal benchmark specifically designed for Chinese Classical Studies across multiple subdomains to bridge this gap. |
| Outcome: | The proposed benchmark spans seven core subdomains with a total of 45 meticulously designed tasks. |
Fuzzy Reasoning Chain (FRC): An Innovative Reasoning Framework from Fuzziness to Clarity (2025.findings-emnlp)
Copied to clipboard
Ping Chen, Xiang Liu, Zhaoxiang Liu, Zezhou Chen, Xingpeng Zhang, Huan Hu, Zipeng Wang, Kai Wang, Shuming Shi, Shiguo Lian
| Challenge: | ambiguity, polysemy, or uncertainty remain significant challenges in natural language processing. |
| Approach: | They introduce a framework that integrates LLM semantic priors with continuous fuzzy membership degrees to create an explicit interaction between probability-based reasoning and fuzzy membership reasoning. |
| Outcome: | The proposed framework integrates semantic priors with continuous fuzzy membership degrees . it allows ambiguous inputs to be gradually transformed into clear and interpretable decisions . |