Papers by Yao Zhao

The Elephant in the Room: Exploring the Role of Neutral Words in Language Model Group-Agnostic Debiasing (2025.findings-acl)

Challenge:	Existing methods of acceleration require fine-tuning of considerably large models, such as Llama-7B, posing a challenge for average users.
Approach:	They propose a Guidance-based Knowledge Transfer framework that leverages a larger LLM as a 'teacher' and a smaller 'student' model to finalize responses.
Outcome:	The proposed framework achieves a maximum accuracy improvement of 14.18%, along with a 10.72 times speed-up on GSM8K and an accuracy improvement 14.00% along with 7.73 times speed up in CSQA.

Copied to clipboard

Xinwei Guo, Jiashi Gao, Junlei Zhou, Jiaxin Zhang, Guanhua Chen, Xiangyu Zhao, Quanying Liu, Haiyan Wu, Xin Yao, Xuetao Wei

Layer-wise Importance Matters: Less Memory for Better Performance in Parameter-efficient Fine-tuning of Large Language Models (2024.findings-emnlp)

Challenge:	Large Language Models (LLMs) are increasingly integrated into our daily lives, raising ethical concerns, especially about perpetuating stereotypes.
Approach:	They propose a method that incorporates a neutral word semantics-based loss function to alleviate the deterioration of the LMS during debiasing.
Outcome:	The proposed method alleviates the deterioration of the Language Modeling Score (LMS) by incorporating a neutral word semantics-based loss function.

Copied to clipboard

Kai Yao, Penglei Gao, Lichun Li, Yuan Zhao, Xiaofeng Wang, Wei Wang, Jianke Zhu

MedQA-CS: Objective Structured Clinical Examination (OSCE)-Style Benchmark for Evaluating LLM Clinical Skills (2026.eacl-long)

Challenge:	Parameter-Efficient Fine-Tuning (PEFT) methods have gained popularity for adapting pre-trained Large Language Models (LLMs) to downstream tasks.
Approach:	They propose a method to optimize the importance of full layers with layer-wise importance scoring by leveraging the estimated importance scores.
Outcome:	The proposed method is compatible with PEFT methods that operate on a per-layer basis and achieves better performance.

Copied to clipboard

Zonghai Yao, Zihao Zhang, Chaolong Tang, Xingyu Bian, Youxia Zhao, Zhichao Yang, Junda Wang, Huixue Zhou, Won Seok Jang, Feiyun Ouyang, Hong Yu

Challenge:	Current clinical LLM benchmarks fail to evaluate advanced clinical skills in AI and large language models (LLMs).
Approach:	They propose a framework to evaluate large language models (LLMs) using two instruction-following tasks designed to reflect real clinical scenarios.
Outcome:	The proposed framework evaluates LLMs through two instruction-following tasks designed to reflect real clinical scenarios.

Generate-on-Graph: Treat LLM as both Agent and KG for Incomplete Knowledge Graph Question Answering (2024.emnlp-main)

Copied to clipboard

Yao Xu, Shizhu He, Jiabei Chen, Zihao Wang, Yangqiu Song, Hanghang Tong, Guang Liu, Jun Zhao, Kang Liu

Structured Data in NLP

Challenge:	Existing methods to integrate LLMs with Knowledge Graphs (KGs) however, these methods are often incomplete to cover all the knowledge required to answer questions.
Approach:	They propose to integrate LLMs with Knowledge Graphs (KGs) to address insufficient knowledge and hallucination issues in Large Language Models.
Outcome:	The proposed method outperforms existing methods on two datasets.

Fact-and-Reflection (FaR) Improves Confidence Calibration of Large Language Models (2024.findings-acl)

Copied to clipboard

Xinran Zhao, Hongming Zhang, Xiaoman Pan, Wenlin Yao, Dong Yu, Tongshuang Wu, Jianshu Chen

ForumSum: A Multi-Speaker Conversation Summarization Dataset (2021.findings-emnlp)

Challenge:	Existing studies on the confidence calibration of LLMs have not explored the effects of different prompting strategies on LLM performance.
Approach:	They propose Fact-and-Reflection prompting which improves LLM confidence calibration . they propose to use human cognition to elicit known "facts" and ask model to "reflect" over them .
Outcome:	The proposed method lowers the expected calibration error by 23.5% on multi-purpose QA tasks.

Copied to clipboard

Misha Khalman, Yao Zhao, Mohammad Saleh

Information Extraction & Text Mining

Challenge:	Abstractive summarization quality has been improved but there is a lack of data for conversation summarizing applications.
Approach:	They propose to build a conversation summarization dataset with human written summaries from internet forums.
Outcome:	The proposed dataset can be easily expanded to improve conversation summarization applications.

Query2Triple: Unified Query Encoding for Answering Diverse Complex Queries over Knowledge Graphs (2023.findings-emnlp)

Copied to clipboard

Yao Xu, Shizhu He, Cunguang Wang, Li Cai, Kang Liu, Jun Zhao

Data Augmentation for Few-Shot Knowledge Graph Completion from Hierarchical Perspective (2022.coling-1)

Challenge:	Complex Query Answering (CQA) is a challenge task of Knowledge Graphs due to incompleteness of KGs.
Approach:	They propose a query embedding approach that decouples the training for simple and complex queries.
Outcome:	The proposed approach decouples training for simple and complex queries and achieves state-of-the-art performance over three public benchmarks.

Copied to clipboard

Yuanzhou Yao, Zhao Zhang, Yongjun Xu, Chao Li

An Auxiliary Task Boosted Multi-task Learning Method for Service Account Retrieval with Limited Human Annotation (2023.emnlp-industry)

Challenge:	Existing knowledge graph completion models require only a few associative triples to complete a relationship.
Approach:	They propose to perform data augmentation from two perspectives to solve the FKGC problem by inferring new triple facts from existing models.
Outcome:	The proposed framework can be applied to a number of existing models.

Copied to clipboard

Yuanzhou Yao, Zhao Zhang, Kaijia Yang, Huasheng Liang, Qiang Yan, Yongjun Xu

Enriching Non-Autoregressive Transformer with Syntactic and Semantic Structures for Neural Machine Translation (2021.eacl-main)

Challenge:	Existing approaches to service account retrieval have limited human annotation, resulting in labor-intensive and time-consuming tasks.
Approach:	They propose an Auxiliary task Boosted Multi-Task Learning method which introduces multiple auxiliary tasks and enhances the performance of the main task, service account retrieval.
Outcome:	The proposed method improves the performance of the main task, service account retrieval.

Copied to clipboard

Ye Liu, Yao Wan, Jianguo Zhang, Wenting Zhao, Philip Yu

Machine Translation

Challenge:	Existing non-autoregressive models have boosted the efficiency of neural machine translation, but their performance is significantly worse than that of autoregressive counterparts.
Approach:	They propose to incorporate syntactic and semantic structures among natural languages into a non-autoregressive Transformer for the task of neural machine translation.
Outcome:	The proposed model achieves faster speed and keeps translation quality compared with other models.

Keyphrase Generation via Soft and Hard Semantic Corrections (2022.emnlp-main)

Copied to clipboard

Guangzhen Zhao, Guoshun Yin, Peng Yang, Yu Yao

R³A: Reinforced Reasoning for Relevance Assessment for RAG in User-Generated Content Platforms (2026.acl-industry)

Challenge:	Extensive experiments show that CorrKG is capable of generating high-quality keyphrases.
Approach:	They propose a correction model CorrKG on top of the MLE pipeline to correct the biases . the adaptive adaptive mass learning scheme is designed to better fit OT and FreqFS .
Outcome:	The proposed model overcomes the semantic biases in keyphrase generation using OT and FreqFS techniques.

Copied to clipboard

Xiaowei Yuan, Lei Jin, Haoxin Zhang, Ziyang Huang, Yan Gao, null Yiwu, Yao Hu, Jun Zhao, Kang Liu

LLMs Trust Humans More, That’s a Problem! Unveiling and Mitigating the Authority Bias in Retrieval-Augmented Generation (2025.acl-long)

Challenge:	Existing approaches to query–document relevance assessment are limited . ambiguous user intent and asymmetric relevance are challenges for RAG platforms .
Approach:	They propose a decomposed reasoning model for relevance assessment that decomposes query intent into intent inference and evidence grounding.
Outcome:	The proposed model outperforms strong baselines on offline benchmarks and achieves significant gains in large-scale online A/B testing.

Copied to clipboard

Yuxuan Li, Xinwei Guo, Jiashi Gao, Guanhua Chen, Xiangyu Zhao, Jiaxin Zhang, Quanying Liu, Haiyan Wu, Xin Yao, Xuetao Wei

Information Retrieval

Challenge:	Large language models (LLMs) generate outputs that stray from user input or contravene established knowledge.
Approach:	They propose a new phenomenon, Authority Bias, where LLMs favor one knowledge source over the other . they propose atomic information that generates conflicts and a Conflict Detection Enhanced Query framework .
Outcome:	The proposed framework reduces Authority bias in large language models . it detects conflicts, performs credibility assessment on conflicting paragraphs, and detects perturbed text .

DIVKNOWQA: Assessing the Reasoning Ability of LLMs via Open-Domain Question Answering over Knowledge Base and Text (2024.findings-naacl)

Copied to clipboard

Wenting Zhao, Ye Liu, Tong Niu, Yao Wan, Philip Yu, Shafiq Joty, Yingbo Zhou, Semih Yavuz

Information Retrieval

Challenge:	Retrievalaugmented LLMs have been used to ground LLM in external knowledge . a gap exists in the current landscape regarding the effectiveness of grounding LLM on heterogeneous knowledge sources.
Approach:	They propose a model that uses symbolic language to generate symbolic queries . they use a dataset that is generated using predefined reasoning chains and human annotation .
Outcome:	The proposed model outperforms previous approaches by a significant margin in QA tasks over text.

Robust Tool Use via Fission-GRPO: Learning to Recover from Execution Errors (2026.acl-long)

Copied to clipboard

Zhiwei Zhang, Fei Zhao, Rui Wang, Zezhong Wang, Bin Liang, Jiakang Wang, Yao Hu, Shaosheng Cao, Kam-Fai Wong

kNN-ICL: Compositional Task-Oriented Parsing Generalization with Nearest Neighbor In-Context Learning (2024.naacl-long)

Challenge:	Large language models (LLMs) can call tools effectively, but they remain brittle in multi-turn execution.
Approach:	They propose a framework that converts execution errors into on-policy corrective supervision within the RL training loop.
Outcome:	The proposed framework improves the error recovery rate of Qwen3-8B by 5.7% absolute and overall accuracy by 4.0% on BFCL v4 Multi-Turn.

Copied to clipboard

Wenting Zhao, Ye Liu, Yao Wan, Yibo Wang, Qingyang Wu, Zhongfen Deng, Jiangshu Du, Shuaiqi Liu, Yunlong Xu, Philip Yu

Large Language Models are In-context Teachers for Knowledge Reasoning (2024.findings-emnlp)

Challenge:	Recent advances in task-oriented parsing involve formulating the task as a sequence-to-sequence problem, relying on a wealth of labeled data.
Approach:	They propose a task-oriented parsing framework that integrates nearest-neighbor learning with a nearest-nearest approach.
Outcome:	The proposed model can be used to synthesize computer programs based on a natural-language prompt without additional data or specialized prompts.

Copied to clipboard

Jiachen Zhao, Zonghai Yao, Zhichao Yang, Hong Yu

SciAssess: Benchmarking LLM Proficiency in Scientific Literature Analysis (2025.findings-naacl)

Challenge:	In-context teaching is a method of providing in-concept example rationales to a student to reason over unseen cases.
Approach:	They propose to use an LLM's self-elicited explanations as in-context demonstrations to prompt a student to reason over unseen cases.
Outcome:	The proposed model outperforms human-crafted demonstrations on medical question answering and human-created models outperfect human-made demonstrations.

Copied to clipboard

Hengxing Cai, Xiaochen Cai, Junhan Chang, Sihang Li, Lin Yao, Wang Changxin, Zhifeng Gao, Hongshuai Wang, Li Yongge, Mujie Lin, Shuwen Yang, Jiankun Wang, Mingjun Xu, Jin Huang, Xi Fang, Jiaxi Zhuang, Yuqi Yin, Yaqi Li, Changhong Chen, Zheng Cheng, Zifeng Zhao, Linfeng Zhang, Guolin Ke

Challenge:	Existing benchmarks fail to adequately evaluate the proficiency of Large Language Models (LLMs) Existing standards do not cover the skills needed to evaluate LLMs in scientific literature analysis.
Approach:	They propose a benchmark to evaluate the proficiency of large language models in scientific literature analysis.
Outcome:	SciAssess evaluates 11 LLMs on multiple tasks across scientific fields.

Safety is Not Only About Refusal: Reasoning-Enhanced Fine-tuning for Interpretable LLM Safety (2025.findings-acl)

Copied to clipboard

Yuyou Zhang, Miao Li, William Han, Yihang Yao, Zhepeng Cen, Ding Zhao

Mitigating the Alignment Tax of RLHF (2024.emnlp-main)

Challenge:	Large Language Models (LLMs) are vulnerable to jailbreak attacks that exploit weaknesses in traditional safety alignment.
Approach:	They propose a framework that trains models to engage in explicit safe reasoning before response . they propose RATIONAL, which allows models to reject harmful prompts while providing meaningful and context-aware responses.
Outcome:	The proposed framework fine-tunes models to reason about query intent, ethics, and potential harm.

Copied to clipboard

Yong Lin, Hangyu Lin, Wei Xiong, Shizhe Diao, Jianmeng Liu, Jipeng Zhang, Rui Pan, Haoxiang Wang, Wenbin Hu, Hanning Zhang, Hanze Dong, Renjie Pi, Han Zhao, Nan Jiang, Heng Ji, Yuan Yao, Tong Zhang

Factuality Assessment as Modal Dependency Parsing (2021.acl-long)

Challenge:	Large Language Models (LLMs) acquire a wide range of abilities during pre-training, but aligning LLMs under Reinforcement Learning with Human Feedback (RLHF) can lead to forgetting pretrained abilities, which is also known as the alignment tax.
Approach:	They propose to use a model averaging technique to find the most powerful alignment-forging Pareto front among RLHF algorithms.
Outcome:	The proposed method achieves the strongest alignment-forging Pareto front among competing methods.

Copied to clipboard

Jiarui Yao, Haoling Qiu, Jin Zhao, Bonan Min, Nianwen Xue

Syntactic Text Processing

Challenge:	a critical step towards factuality assessment is to determine the factuality of events in text.
Approach:	They propose a modal dependency parsing task that assesses the factuality of events in text . they crowdsource a large-scale data set annotated with modal dependence structures .
Outcome:	The proposed model outperforms the pipeline model in factuality assessment . the proposed model is based on a crowdsourced dataset .

Improving Alignment in LVLMs with Debiased Self-Judgment (2025.findings-emnlp)

Copied to clipboard

Sihan Yang, Chenhang Cui, Zihao Zhao, Yiyang Zhou, Weilong Yan, Ying Wei, Huaxiu Yao

Seeking Rational Demonstrations for Large Language Models: A Domain Generalization Approach to Unsupervised Cross-Domain Keyphrase Generation (2025.acl-short)

Challenge:	Existing methods for aligning LVLMs rely on external datasets, human annotations or complex post-processing.
Approach:	They propose a method that generates a debiased self-judgment score for LVLMs . this self-evaluation metric is created internally by the model without external resources .
Outcome:	The proposed approach outperforms existing methods in reducing hallucinations and safety concerns.

Copied to clipboard

Guangzhen Zhao, Yu Yao, Dechang Kong, Zhenjiang Dong

Metacognitive Self-Correction for Multi-Agent System via Prototype-Guided Next-Execution Reconstruction (2026.findings-acl)

Challenge:	Unsupervised cross-domain keyphrase generation is crucial in real-world natural language processing scenarios, but its accuracy is limited by the distribution shift between source and target domain.
Approach:	They propose to seek rational demonstrations from the source domain and to use them to improve their ability in the unsupervised cross-domain keyphrase generation setting.
Outcome:	The proposed model achieves state-of-the-art on widely used cross-domain KG benchmarks and the results are published in the journal Nature.

Copied to clipboard

Xu Shen, Qi Zhang, Song Wang, Zhen Tan, Xinyu Zhao, Laura Yao, Vaishnav Tadiparthi, Hossein Nourkhiz Mahjoub, Ehsan Moradi Pari, Kwonjoon Lee, Tianlong Chen

PyramidInfer: Pyramid KV Cache Compression for High-throughput LLM Inference (2024.findings-acl)

Challenge:	Large Language Model based multi-agent systems (MAS) excel at collaborative problem solving but remain brittle to cascading errors.
Approach:	They propose a metacognitive framework that enables step-level error detection and self-correction in Large Language Model based multi-agent systems (MAS) .
Outcome:	The proposed framework outperforms baselines on the Who When benchmark and delivers consistent gains on AgentErrorBench.

Copied to clipboard

Dongjie Yang, Xiaodong Han, Yan Gao, Yao Hu, Shilin Zhang, Hai Zhao

Challenge:	Existing methods to reduce memory usage for large language models neglect inter-layer dependency between layers and huge memory consumption in pre-computation.
Approach:	They propose a method that compresses the KV cache by layer-wise retaining crucial context.
Outcome:	The proposed method reduces memory usage by layer-wise retaining crucial context . it can improve 2.2x throughput compared to Accelerate with over 54% memory reduction .

MathSight: A Benchmark Exploring Have Vision-Language Models Really Seen in University-Level Mathematical Reasoning? (2026.acl-long)

Copied to clipboard

Yuandong Wang, Yao Cui, Yuxin Zhao, Zhen Yang, Yangfu Zhu, Zhenzhou Shao

A Survey on Sparse Autoencoders: Interpreting the Internal Mechanisms of Large Language Models (2025.findings-emnlp)

Challenge:	Existing benchmarks rarely isolate how much visual information contributes to reasoning . a growing collection of benchmarks has catalyzed rapid progress in multimodal reasoning - but how much it contributes remains unclear .
Approach:	They propose a university-level multimodal mathematical reasoning benchmark to quantify the effect of visual input.
Outcome:	The proposed benchmark disentangles and quantifies the effect of visual input on multimodal reasoning models.

Copied to clipboard

Dong Shu, Xuansheng Wu, Haiyan Zhao, Daking Rai, Ziyu Yao, Ninghao Liu, Mengnan Du

Explainability & Interpretability in NLP

Challenge:	Sparse Autoencoders (SAEs) can disentangle complex features into more interpretable components.
Approach:	They propose to use Sparse Autoencoders to disentangle LLM features into more interpretable components.
Outcome:	The proposed method disentangles complex features into more interpretable components.

DIDS: Domain Impact-aware Data Sampling for Large Language Model Training (2025.emnlp-main)

Copied to clipboard

Weijie Shi, Jipeng Zhang, Yaguang Wu, Jingzhi Fang, Shibo Zhang, Yao Zhao, Hao Chen, Ruiyuan Zhang, Yue Cui, Jia Zhu, Sirui Han, Jiajie Xu, Xiaofang Zhou

Improving Meta-learning for Low-resource Text Classification and Generation via Memory Imitation (2022.acl-long)

Challenge:	Existing approaches for optimizing domain-level sampling strategies struggle with maintaining intra-domain consistency and accurately measuring domain impact.
Approach:	They propose to use a Fisher-Information Matrix-guided metric to measure domain impact to ensure intra-domain consistency and accuracy.
Outcome:	The proposed model achieves 3.4% higher average performance while maintaining comparable training efficiency.

Copied to clipboard

Yingxiu Zhao, Zhiliang Tian, Huaxiu Yao, Yinhe Zheng, Dongkyu Lee, Yiping Song, Jian Sun, Nevin Zhang

Information Extraction & Text Mining

Challenge:	Building models of natural language processing (NLP) is challenging in low-resource scenarios where limited data are available.
Approach:	They propose a memory imitation meta-learning method that enhances the model’s reliance on support sets for task adaptation.
Outcome:	The proposed method outperforms baselines on both text classification and generation tasks.

NarraSum: A Large-Scale Dataset for Abstractive Narrative Summarization (2022.findings-emnlp)

Copied to clipboard

Chao Zhao, Faeze Brahman, Kaiqiang Song, Wenlin Yao, Dian Yu, Snigdha Chaturvedi

Improving the Robustness of Summarization Models by Detecting and Removing Input Noise (2023.findings-emnlp)

Challenge:	Existing studies focus on summarizing news documents or structured documents.
Approach:	They propose to use a large-scale narrative summarization dataset to encourage research . they find there is a performance gap between humans and the models on NarraSum .
Outcome:	The proposed dataset shows that humans and state-of-the-art models perform poorly when summarizing a narrative . it contains 122K narratives collected from synopses of movies and TV episodes with diverse genres .

Copied to clipboard

Kundan Krishna, Yao Zhao, Jie Ren, Balaji Lakshminarayanan, Jiaming Luo, Mohammad Saleh, Peter Liu

MED-COPILOT: A Medical Assistant Powered by GraphRAG and Similar Patient Case Retrieval (2026.acl-demo)

Challenge:	Abstractive summarization models are typically evaluated using test data that is identically distributed as training data.
Approach:	They propose a method to detect and remove input noise from documents to be summarized without extra training or auxiliary models.
Outcome:	The proposed method recovers a large fraction of the loss in performance, sometimes as large as 11 ROUGE-1 points, without extra training, auxiliary models, or prior knowledge of the type of noise.

Copied to clipboard

Shuheng Chen, Namratha Patil, Haonan Pan, Angel Hsing-Chi Hwang, Yao Du, Ruishan Liu, Jieyu Zhao

RealBench: A Chinese Multi-image Understanding Benchmark Close to Real-world Scenarios (2025.findings-emnlp)

Challenge:	MED-COPILOT is an interactive research prototype for evidence-aware clinical reasoning . large language models (LLMs) are prone to hallucinations and lack verifiable evidence grounding .
Approach:	They propose a system that integrates GraphRAG and semantic-keyword similar-patient retrieval to support transparent clinical reasoning.
Outcome:	The proposed system outperforms baseline and standard RAGs on clinical note completion and medical question answering.

Copied to clipboard

Fei Zhao, Chengqiang Lu, Yufan Shen, Qimeng Wang, Yicheng Qian, Haoxin Zhang, Yan Gao, null Yiwu, Yao Hu, Zhen Wu, Shangyu Xing, Xinyu Dai

Low-Rank Interconnected Adaptation across Layers (2025.findings-acl)

Challenge:	RealBench is the first Chinese multimodal multi-image dataset . the dataset contains 9393 samples and 69910 images .
Approach:	They propose to create a Chinese multimodal multi-image dataset using 21 models . they use closed-source models that support multi-inputs as well as open-source visual and video models a .
Outcome:	The first Chinese multimodal multi-image dataset contains 9393 samples and 69910 images.

Copied to clipboard

Yibo Zhong, Jinman Zhao, Yao Zhou

Green & Sustainable NLP

Challenge:	Low-rank adaptation (LoRA) is a parameter-efficient fine-tuning (PEFT) method that learns weight updates W = AB for pretrained weights W through low-rank adapters A and B.
Approach:	They propose a low-rank interconnected adaptation across layers method that introduces an interconnected framework with locally shared A and globally shared B experts.
Outcome:	The proposed method improves expressiveness across domains and modalities and enables higher-rank W with equal or fewer parameters.

LLaSA: Large Language and Structured Data Assistant (2025.naacl-long)

Copied to clipboard

Yao Xu, Shizhu He, Jiabei Chen, ZengXiangrong ZengXiangrong, Bingning Wang, Guang Liu, Jun Zhao, Kang Liu

GradOT: Training-free Gradient-preserving Offsite-tuning for Large Language Models (2025.acl-long)

Challenge:	Structured knowledge grounding (SKG) tasks are a key part of many NLP applications.
Approach:	They propose a framework for enhancing LLMs' ability to handle structured data . they represent various types of structured data in a unified hypergraph format .
Outcome:	The proposed framework outperforms existing methods on SKG tasks using LoRA finetuning.

Copied to clipboard

Kai Yao, Zhaorui Tan, Penglei Gao, Lichun Li, Kaixin Wu, Yinggui Wang, Yuan Zhao, Yixin Ji, Jianke Zhu, Wei Wang

MiCRo: Mixture Modeling and Context-aware Routing for Personalized Preference Learning (2025.emnlp-main)

Challenge:	Existing methods for offsite-tuning of large language models require high computational costs and lack theoretical analysis.
Approach:	They propose an offsite-tuning approach that selectively applies compression techniques such as rank compression and channel pruning to preserve the gradients of fine-tuned adapters while ensuring privacy.
Outcome:	The proposed method surpasses existing OT methods in privacy protection and model performance.

Copied to clipboard

Jingyan Shen, Jiarui Yao, Rui Yang, Yifan Sun, Feng Luo, Rui Pan, Tong Zhang, Han Zhao

Awakening Augmented Generation: Learning to Awaken Internal Knowledge of Large Language Models for Question Answering (2025.coling-main)

Challenge:	Existing reward models assume a global reward function, limiting personalization and pluralistic alignment.
Approach:	They propose a framework that leverages binary preference datasets to enhance personalized preference learning.
Outcome:	The proposed framework captures diverse human preferences without fine-grained annotations and significantly improves personalized preference learning on downstream tasks.

Copied to clipboard

Huanxuan Liao, Shizhu He, Yao Xu, Yuanzhe Zhang, Shengping Liu, Kang Liu, Jun Zhao

Personalized Federated Learning for Text Classification with Gradient-Free Prompt Tuning (2024.findings-naacl)

Challenge:	Recent studies indicate that Large Language Models model rich knowledge, but it is often not activated and awakened.
Approach:	They propose a framework that leverages richer context to enhance question answering . Explicit awakening fine-tunes a context generator to create a synthetic, compressed document that functions as symbolic context.
Outcome:	The proposed framework mimics the human ability to answer questions using only thinking and recalling to compensate for knowledge gaps.

Copied to clipboard

Rui Wang, Tong Yu, Ruiyi Zhang, Sungchul Kim, Ryan Rossi, Handong Zhao, Junda Wu, Subrata Mitra, Lina Yao, Ricardo Henao

Your Language Model May Think Too Rigidly: Achieving Reasoning Consistency with Symmetry-Enhanced Training (2025.findings-acl)

Challenge:	Pretrained language models (PLMs) are used for personalized federated learning . communication costs are high with large PLMs, and local training is expensive .
Approach:	They propose a framework for federated learning with pretrained language models . they propose 'discrete local search' and compression mechanism for local training .
Outcome:	The proposed framework achieves superior performance compared with baselines.

Copied to clipboard

Yihang Yao, Zhepeng Cen, Miao Li, William Han, Yuyou Zhang, Emerson Liu, Zuxin Liu, Chuang Gan, Ding Zhao

Can LLMs Hear the Dogwhistle? (2026.findings-acl)

Challenge:	Large Language Models (LLMs) have demonstrated strong reasoning capabilities across various tasks.
Approach:	They propose a data-centric approach that enhances LLMs’ awareness of symmetry in query variations and propose syMmetry-ENhanceD (MEND) data augmentation.
Outcome:	Extensive experiments on logical and arithmetic reasoning tasks show that the proposed approach improves model robustness at the knowledge extraction stage through query augmentation.

Copied to clipboard

Yifan Liu, Yi Lin, Xinwei Guo, Ziwei Wang, Jiaxin Zhang, Guanhua Chen, Haiyan Wu, Xiangyu Zhao, Xin Yao, Xuetao Wei

GCPG: A General Framework for Controllable Paraphrase Generation (2022.findings-acl)

Challenge:	Existing safety benchmarks focus on explicitly harmful content, but ignore context-dependent expressions such as dogwhistles.
Approach:	They propose a benchmark for evaluating LLM safety under dogwhistle-driven prompts . their findings expose a blind spot in current safety evaluation practices .
Outcome:	The proposed benchmark compared safety performance with toxic terms using dogwhistle-driven prompts.

Copied to clipboard

Kexin Yang, Dayiheng Liu, Wenqiang Lei, Baosong Yang, Haibo Zhang, Xue Zhao, Wenqing Yao, Boxing Chen

Planning with Learned Entity Prompts for Abstractive Summarization (2021.tacl-1)

Challenge:	Existing studies highlight a special condition under two indispensable aspects of controllable paraphrase generation (CPG) individually, lacking a unified circumstance to explore and analyze their effectiveness.
Approach:	They propose a general controllable paraphrase generation framework that integrates lexical and syntactical conditions into a text sequence and uniformly processes them in an encoder-decoder paradigm.
Outcome:	The proposed framework can combine lexical and syntactical conditions and improve paraphrase generation.

Copied to clipboard

Shashi Narayan, Yao Zhao, Joshua Maynez, Gonçalo Simões, Vitaly Nikolaev, Ryan McDonald

PRIME: A Process-Outcome Alignment Benchmark for Verifiable Reasoning in Mathematics and Engineering (2026.acl-long)

Challenge:	a simple but flexible mechanism is used to ground the generation of abstractive summaries.
Approach:	They propose a mechanism to learn an intermediate plan to ground the generation of abstractive summaries.
Outcome:	The proposed model outperforms state-of-the-art methods for faithfulness on CNN and BillSum.

Copied to clipboard

Xiangfeng Wang, Hangyu Guo, Yanlin Lai, Mitt Huang, Liang Zhao, Chengyuan Yao, Yinmin Zhang, Qi Han, null Xiaoxiaoren, Chun Yuan, Tong Xu, Zheng Ge, Xiangyu Zhang, Daxin Jiang

Enhancing LLM-based Hatred and Toxicity Detection with Meta-Toxic Knowledge Graph (2025.findings-acl)

Challenge:	Current outcome-centric verification paradigms neglect potential errors in the derivation process.
Approach:	They propose a process-aware RLVR training paradigm utilizing verifiers selected via PRIME.
Outcome:	The proposed approach outperforms the baseline verification paradigm on AIME24, AIME25, and Beyond-AIME models.

Copied to clipboard

Yibo Zhao, Jiapeng Zhu, Can Xu, Yao Liu, Xiang Li

AutoPRM: Automating Procedural Supervision for Multi-Step Reasoning via Controllable Question Decomposition (2024.naacl-long)

Challenge:	Existing methods to address toxicity issues with large language models are inadequate . lack of domain-specific knowledge leads to false negatives and excessive sensitivity to toxic speech limits freedom of speech.
Approach:	They propose a method that leverages graph search on a meta-toxic knowledge graph to enhance hatred and toxicity detection.
Outcome:	The proposed method lowers false positive rate and improves toxicity detection performance in out-of-domain scenarios.

Copied to clipboard

Zhaorun Chen, Zhuokai Zhao, Zhihong Zhu, Ruiqi Zhang, Xiang Li, Bhiksha Raj, Huaxiu Yao

Tailored Primitive Initialization is the Secret Key to Reinforcement Learning (2026.acl-long)

Challenge:	Recent advances in large language models (LLMs) have shown promise in multi-step reasoning tasks, yet relying on extensive manual labeling to provide procedural feedback remains a significant impediment.
Approach:	They propose a self-supervised framework that decomposes complex problems into manageable subquestions with a controllable granularity switch and sequentially applies reinforcement learning to iteratively improve the subquest solver.
Outcome:	The proposed framework improves performance on mathematical and commonsense reasoning tasks over SOTA.

Copied to clipboard

Yihang Yao, Guangtao Zeng, Raina Wu, Yang Zhang, Ding Zhao, Zhang-Wei Hong, Chuang Gan

ChildTalk: A Multi-Dialect Chinese Child Speech Corpus with Full-Length Child–Caregiver Conversations for Speech Recognition (2026.findings-acl)

Challenge:	Reinforcement learning (RL) has emerged as a powerful paradigm for improving the reasoning capabilities of large language models.
Approach:	They propose a pipeline that automatically discovers thinking token patterns with reasoning primitives and curates SFT datasets to prepare LLMs for RL.
Outcome:	The proposed pipeline outperforms baseline methods on mathematical and logical reasoning benchmarks on RL tasks.

Copied to clipboard

Jiaming Zhou, Yujie Guo, Shiwan Zhao, Yao Lu, Jianye Wang, Haoqin Sun, Hui Wang, Yong Qin

Dialogue Systems & Conversational Agents

Challenge:	Automatic speech recognition (ASR) for children remains challenging due to developmental variability and the scarcity of high-quality corpora.
Approach:	They propose a large-scale Chinese child speech corpus that contains 112.5 hours of speech from 498 children and 500 caregivers.
Outcome:	The proposed model improves in-domain and cross-domain performance on children's speech.

Attend, Memorize and Generate: Towards Faithful Table-to-Text Generation in Few Shots (2021.findings-emnlp)

Copied to clipboard

Wenting Zhao, Ye Liu, Yao Wan, Philip Yu

XQuant: Achieving Ultra-Low Bit KV Cache Quantization with Cross-Layer Compression (2025.emnlp-main)

Challenge:	Few-shot table-to-text generation is a task of composing fluent and faithful sentences to convey table content using limited data.
Approach:	They propose a novel approach to generate faithful table-to-text sentences using limited data . they aim to exploit table structure and natural linguistic information to generate accurate sentences .
Outcome:	The proposed approach generates higher qualified sentences when compared with state-of-the-art models on humans, songs, and books.

Copied to clipboard

Haoqi Yang, Yao Yao, Zuchao Li, Baoyuan Qi, Liu Guoming, Hai Zhao

Challenge:	Large Language Models (LLMs) have demonstrated remarkable capabilities across diverse tasks. however, their extensive memory requirements present significant challenges for deployment in resource-constrained environments.
Approach:	They propose a training-free framework that achieves ultra-low equivalent bit-width KV cache quantization.
Outcome:	The proposed framework outperforms state-of-the-art methods on TruthfulQA and LongBench.

Personalizing LLMs with Binary Feedback: A Preference-Calibrated Optimization Framework (2026.acl-long)

Copied to clipboard

Xilai Ma, Liye Zhao, Weijun Yao, Haibing Di, Wenya Wang, Jing Li

Challenge:	Existing methods focus on isolated user histories, neglecting the essential role of inter-user differences.
Approach:	They propose a framework that personalizes Large Language Models via preference-calibrated binary signals.
Outcome:	The proposed framework outperforms baselines in a variety of personalization tasks and backbone LLMs.

RU22Fact: Optimizing Evidence for Multilingual Explainable Fact-Checking on Russia-Ukraine Conflict (2024.lrec-main)

Copied to clipboard

Yirong Zeng, Xiao Ding, Yi Zhao, Xiangyu Li, Jie Zhang, Chao Yao, Ting Liu, Bing Qin

GLIMPSE: Do Large Vision-Language Models Truly Think With Videos or Just Glimpse at Them? (2025.emnlp-main)

Challenge:	Existing methods to verify factuality of claims do not provide sufficient evidence for explainable fact-checking systems.
Approach:	They propose a method to automatically retrieve and summarize evidence from the Web and a novel multilingual explainable fact-checking dataset on the Russia-Ukraine conflict in 2022.
Outcome:	The proposed method can retrieve and summarize evidence from the Web and generate explanations in 16 languages.

Copied to clipboard

Yiyang Zhou, Linjie Li, Shi Qiu, Zhengyuan Yang, Yuyang Zhao, Siwei Han, Yangfan He, Kangqi Li, Haonian Ji, Zihao Zhao, Haibo Tong, Lijuan Wang, Huaxiu Yao

Investigating Ensemble Methods for Model Robustness Improvement of Text Classifiers (2022.findings-emnlp)

Challenge:	Existing video benchmarks often resemble image-based questions with scans of only a few key frames, without deep temporal reasoning.
Approach:	They propose a video benchmark to assess whether large vision-language models can genuinely think with videos rather than perform superficial frame-level analysis.
Outcome:	The proposed benchmark consists of 3,269 videos and over 4,342 highly visual-centric questions across 11 categories, including Trajectory Analysis, Temporal Reasoning, and Forensics Detection.

Copied to clipboard

Jieyu Zhao, Xuezhi Wang, Yao Qin, Jilin Chen, Kai-Wei Chang

Sign2Vis: Automated Data Visualization from Sign Language (2025.findings-acl)

Challenge:	Existing methods to reduce model's reliance on bias features ignore the learnability of these features.
Approach:	They propose to reduce models' reliance on bias features by first training models with fixed low-capacity models which ignore the learnability of the bias features.
Outcome:	The proposed models can perform better on out-of-distribution datasets than baseline models with a more sophisticated model design.

Copied to clipboard

Yao Wan, Yang Wu, Zhen Li, Guobiao Zhang, Hongyu Zhang, Zhou Zhao, Hai Jin, April Wang

Challenge:	Existing methods to translate natural language descriptions into visualization queries focus on spoken languages, not sign languages.
Approach:	They propose a sign language interface that enables the DHH community to engage more fully with data analysis.
Outcome:	The proposed interface can be used by the deaf and hard-of-hearing community.

LiPO: Listwise Preference Optimization through Learning-to-Rank (2025.naacl-long)

Copied to clipboard

Tianqi Liu, Zhen Qin, Junru Wu, Jiaming Shen, Misha Khalman, Rishabh Joshi, Yao Zhao, Mohammad Saleh, Simon Baumgartner, Jialu Liu, Peter J Liu, Xuanhui Wang

Paragraph-level Neural Question Generation with Maxout Pointer and Gated Self-attention Networks (D18-1)

Challenge:	Recent work on language models with curated feedback provides promising alternatives to RLHF . multiple responses can be ranked by reward models or AI feedback, but there is no study on directly fitting upon a list of responses.
Approach:	They propose a method that aligns language models with curated human feedback . they propose SLiC and DPO as promising alternatives to traditional RLHF .
Outcome:	The proposed method outperforms DPO and SLiC on several preference alignment tasks with curated and real rankwise preference data.

Copied to clipboard

Yao Zhao, Xiaochuan Ni, Yuanyuan Ding, Qifa Ke

Illusions of Confidence? Diagnosing LLM Truthfulness via Neighborhood Consistency (2026.acl-long)

Challenge:	Existing rule-based question generation models rely on one or two sentences as input, while long text has posed challenges for sequence to sequence neural models.
Approach:	They propose a maxout pointer mechanism with gated self-attention encoder to address the challenges of processing long text inputs for question generation.
Outcome:	The proposed model outperforms existing models with sentence-level or paragraph-level inputs pushing the state-of-the-art result from 13.9 to 16.3 (BLEU_4).

Copied to clipboard

Haoming Xu, Ningyuan Zhao, Yunzhi Yao, Weihong Xu, Hongru Wang, Xinle Deng, Shumin Deng, Jeff Z. Pan, Huajun Chen, Ningyu Zhang

Challenge:	Existing evaluations rely on point-wise confidence, which can mask brittle belief.
Approach:	They propose a measure of belief robustness that evaluates coherence across a conceptual neighborhood.
Outcome:	The proposed model is more resistant to interference than existing models.

Reframing Responsibility: Framing-Aware Event Causality Identification (2026.acl-long)

Copied to clipboard

Jin Zhao, Jiayi Yao, Xinrui Hu, Nianwen Xue

Learning-by-Narrating: Narrative Pre-Training for Zero-Shot Dialogue Comprehension (2022.acl-short)

Challenge:	Causal explanations in political narratives are often framed and contested.
Approach:	They propose a framing-aware extension of ECI that models causal explanations as structured claims including responsibility targets, evaluative frams, source type, and epistemic modality.
Outcome:	The proposed model enables quantitative analysis of divergent causal attribution across narratives.

Copied to clipboard

Chao Zhao, Wenlin Yao, Dian Yu, Kaiqiang Song, Dong Yu, Jianshu Chen

Dialogue Systems & Conversational Agents

Challenge:	Existing models for dialogue comprehension are not available for the pre-training of such a model.
Approach:	They propose a narrative-guided pre-training strategy that learns by narrating key information from a dialogue input.
Outcome:	The proposed model performs better on four dialogue-based tasks and is comparable to existing models.

GUICourse: From General Vision Language Model to Versatile GUI Agent (2025.acl-long)

Copied to clipboard

Wentong Chen, Junbo Cui, Jinyi Hu, Yujia Qin, Junjie Fang, Yue Zhao, Chongyi Wang, Jun Liu, Guirong Chen, Yupeng Huo, Yuan Yao, Yankai Lin, Zhiyuan Liu, Maosong Sun

Visual Data in NLP

Challenge:	Graphical User Interfaces (GUIs) are a pivotal medium for human-computer interaction.
Approach:	They propose a series of datasets for training visual-based GUI agents using general VLMs.
Outcome:	The proposed GUICourse datasets show that even a small-sized GUI agent performs better on GUI tasks.

BoYaEval: Evaluating Multimodal Large Language Models on Understanding Ancient Chinese Musical Scores (2026.acl-long)

Copied to clipboard

Jiajia Li, Weizhi Xue, Yao Yao, Qiwei Li, null Chenchong, Zuchao Li, Ping Wang, Hai Zhao

SWE-Mutation: Can LLMs Generate Reliable Test Suites in Software Engineering? (2026.findings-acl)

Challenge:	Multimodal Large Language Models excel in general tasks but struggle with specialized, structured cultural symbols.
Approach:	They evaluate 21 leading MLLMs and compare their performance to a benchmark for Ancient Chinese musical notation.
Outcome:	The benchmark evaluates 21 leading MLLMs on five types of ancient Chinese music notation systems.

Copied to clipboard

Yuxuan Sun, Yuze Zhao, Yufeng Wang, Yao Du, Zhiyuan Ma, Jinbo Wang, Mengdi Zhang, Kai Zhang, Zhenya Huang

Implicit Relation Linking for Question Answering over Knowledge Graph (2022.findings-acl)

Challenge:	Evaluating software engineering capabilities is a core component of large language models (LLMs).
Approach:	They propose a benchmark to evaluate LLM-generated test suites that introduces mutated solutions that attempt to "fool" them.
Outcome:	The proposed test suites are based on 2,636 mutated variants derived from 800 original instances and include a multilingual subset spanning nine programming languages.

Copied to clipboard

Yao Zhao, Jiacheng Huang, Wei Hu, Qijin Chen, XiaoXia Qiu, Chengfu Huo, Weijun Ren

EICO: Improving Few-Shot Text Classification via Explicit and Implicit Consistency Regularization (2022.findings-acl)

Challenge:	Existing methods rely on textual similarities between NL and KG to build relation links.
Approach:	They propose an implicit relation linking method called ImRL which links relation phrases in NL to relation paths in KG.
Outcome:	The proposed method significantly outperforms state-of-the-art methods on two benchmarks and a newly-created datasets.

Copied to clipboard

Lei Zhao, Cheng Yao

TCSinger 2: Customizable Multilingual Zero-shot Singing Voice Synthesis (2025.findings-acl)

Challenge:	Existing methods for few-shot text classification are limited by labeled data.
Approach:	They propose to use consistency regularization to improve few-shot text classification by generating pseudo-labels from weakly-augmented and strongly-augmented views.
Outcome:	The proposed method achieves competitive performance with 16 labeled examples with prompt and verbalizer.

Copied to clipboard

Yu Zhang, Wenxiang Guo, Changhao Pan, Dongyu Yao, Zhiyuan Zhu, Ziyue Jiang, Yuhan Wang, Tao Jin, Zhou Zhao

MPBench: A Comprehensive Multimodal Reasoning Benchmark for Process Errors Identification (2025.findings-acl)

Challenge:	Existing zero-shot singing voice synthesis models depend on phoneme and note boundary annotations, limiting their robustness and producing poor transitions between phonemes and notes.
Approach:	They propose a multi-task multilingual zero-shot SVS model with style transfer and style control based on various prompts.
Outcome:	Experimental results show that TCSinger 2 outperforms baseline models in subjective and objective metrics across multiple related tasks.

Copied to clipboard

xu Zhao Pan, Pengfei Zhou, Jiaxin Ai, Wangbo Zhao, Kai Wang, Xiaojiang Peng, Wenqi Shao, Hongxun Yao, Kaipeng Zhang

Beyond Scaling: Predicting Patent Approval with Domain-specific Fine-grained Claim Dependency Graph (2024.acl-long)

Challenge:	Existing benchmarks of large language models focus on error detection, neglecting other scenarios like reasoning search.
Approach:	et al. propose a multi-task, multimodal benchmark to assess effectiveness of PRMs . step correctness, answers aggregation and reasoning process search are evaluated . ethical principles of MPBench are based on a set of evaluation paradigms based in a text-based benchmark .
Outcome:	a new benchmark assesses the effectiveness of large language models (LLMs) in multiple scenarios . it uses three evaluation paradigms to assess the effectiveness and compares them with existing models . a the proposed model improves reasoning accuracy by providing stepwise feedback for multi-step reasoning results .

Copied to clipboard

Xiaochen Gao, Feng Yao, Kewen Zhao, Beilei He, Animesh Kumar, Vish Krishnan, Jingbo Shang

MoDE-CoTD: Chain-of-Thought Distillation for Complex Reasoning Tasks with Mixture of Decoupled LoRA-Experts (2024.lrec-main)

Challenge:	Scaling up language models has demonstrated predictable improvement and unprecedented abilities in many language tasks.
Approach:	They propose a fine-grained cLAim depeNdency graph that captures the dependencies within the patent data and extends the embedding-based state-of-the-art (SOTA) they then explore prompt-based methods to harness proprietary LLMs' potential, but find the best results close to random guessing, underlining the ineffectiveness of model scaling-up.
Outcome:	The proposed graph methods outperform the standard model scaling methods in the patent approval prediction task and show that they are cost-effective.

Copied to clipboard

Xiang Li, Shizhu He, Jiayu Wu, Zhao Yang, Yao Xu, Yang jun Jun, Haifeng Liu, Kang Liu, Jun Zhao

SirLLM: Streaming Infinite Retentive LLM (2024.acl-long)

Challenge:	Current Chain-of-thought Distillation methods hinder CoT reasoning performance . student models are separately distilled from specific reasoning tasks . parameter update of student models severely harms CoT ability on unseen reasoning tasks.
Approach:	They propose a method which distills Chain-of-thought reasoning ability of large language models to much smaller student models.
Outcome:	The proposed method improves the reasoning ability of large language models on 14 datasets.

Copied to clipboard

Yao Yao, Zuchao Li, Hai Zhao

TableLLM: Enabling Tabular Data Manipulation by LLMs in Real Office Usage Scenarios (2025.findings-acl)

Challenge:	Large Language Models (LLMs) are becoming increasingly prevalent in various domains, requiring a one-off input of overly long texts to maintain a degree of memory.
Approach:	They propose a Streaming Infinite Retentive LLM which allows LLMs to maintain longer memory during infinite-length dialogues without fine-tuning.
Outcome:	The proposed model can achieve stable and significant improvements across different LLMs and tasks, compellingly proving its effectiveness.

Copied to clipboard

Xiaokang Zhang, Sijia Luo, Bohan Zhang, Zeyao Ma, Jing Zhang, Yang Li, Guanlin Li, Zijun Yao, Kangli Xu, Jinchang Zhou, Daniel Zhang-Li, Jifan Yu, Shu Zhao, Juanzi Li, Jie Tang

OmniDiagram: Advancing Unified Diagram Code Generation via Visual Interrogation Reward (2026.findings-acl)

Challenge:	TableLLM is a robust large language model capable of handling tabular data manipulation tasks.
Approach:	They propose a distant supervision method for training which includes a reasoning process extension strategy and a cross-way validation strategy.
Outcome:	The proposed model has 8 billion parameters and is capable of handling tabular data tasks.

Copied to clipboard

Haoyue Yang, Xuanle Zhao, Xuexin Liu, Feibing Jiang, Yao Zhu

S2ST-Omni: Hierarchical Language-Aware SpeechLLM Adaptation for Multilingual Speech-to-Speech Translation (2026.findings-acl)

Challenge:	Existing studies on programmable diagram generation focus on a narrow set of tasks and languages.
Approach:	They propose a unified framework that integrates diverse diagram code languages and task definitions.
Outcome:	The proposed framework can bridge complex visual information with executable code across diverse tasks and languages.

Copied to clipboard

Yu Pan, Xiongfei Wu, Yang Yuguang, Jixun Yao, Maxime Cordy, Lei Ma, Jianjun Zhao

Thinking Traps in Long Chain-of-Thought: A Measurable Study and Trap-Aware Adaptive Restart (2026.findings-acl)

Challenge:	S2ST-Omni integrates a speech-to-text frontend with a modular, plug-and-play text-tospeech backend.
Approach:	They propose a compositional S2ST framework that integrates a speech-to-text frontend with a modular, plug-and-play text-tospeech backend.
Outcome:	The proposed framework outperforms existing frameworks in translation and synthesis . it integrates a speech-to-text translation frontend with a plug-and-play text-tospeech backend .

Copied to clipboard

null Chenkang, Fan Yu, Junjie Nian, Sihan Zhao, Zhuoka Feng, Zijun Yao, Wang Heng, Yu Minshen, Yixin Cao

Uncertainty Propagation on LLM Agent (2025.acl-long)

Challenge:	Experiments show that extended generation does not guarantee correctness . a recurring pattern in Long-CoT failures is a problem for large reasoning models .
Approach:	They propose a test-time control framework that truncates the trajectory before the trap segment and adaptively restarts decoding.
Outcome:	Experiments show that TAAR improves reasoning performance without fine-tuning model parameters.

Copied to clipboard

Qiwei Zhao, Dong Li, Yanchi Liu, Wei Cheng, Yiyou Sun, Mika Oishi, Takao Osaki, Katsushi Matsuda, Huaxiu Yao, Chen Zhao, Haifeng Chen, Xujiang Zhao

Challenge:	Existing methods for estimating uncertainty in large language models (LLMs) focus on final-step outputs, which fail to account for cumulative uncertainty over multi-step decision-making process and dynamic interactions between agents and their environments.
Approach:	They propose a framework that propagates uncertainty through each step of an LLM-based agent’s reasoning process.
Outcome:	Extensive experiments on benchmark datasets show that the proposed framework outperforms state-of-the-art methods by 20%.

A Well-Composed Text is Half Done! Composition Sampling for Diverse Conditional Generation (2022.acl-long)

Copied to clipboard

Shashi Narayan, Gonçalo Simões, Yao Zhao, Joshua Maynez, Dipanjan Das, Michael Collins, Mirella Lapata

Investigating Efficiently Extending Transformers for Long Input Summarization (2023.emnlp-main)

Challenge:	Composition Sampling is a simple but effective method to generate diverse outputs for conditional generation of higher quality compared to previous stochastic decoding strategies.
Approach:	They propose a method to generate diverse outputs for conditional generation . they use a plan-based neural generation model that is trained to create a composition of the output and then generate by conditioning on it and the input.
Outcome:	The proposed method avoids text degeneration by first sampling a composition in the form of an entity chain and then using beam search to generate the best possible text grounded to this entity chain.

Copied to clipboard

Jason Phang, Yao Zhao, Peter Liu

Summarization

Challenge:	Large pretrained Transformer models have proven capable at tackling natural language tasks, but handling long sequence inputs still poses a significant challenge.
Approach:	They propose an extension of the PEGASUS model with additional long input pretraining to handle inputs of up to 16K tokens.
Outcome:	The proposed model achieves strong performance on long input summarization tasks comparable with much larger models.

Multi-Modal Generative Adversarial Network for Short Product Title Generation in Mobile E-Commerce (N19-2)

Copied to clipboard

Jianguo Zhang, Pengcheng Zou, Zhao Li, Yao Wan, Xiuming Pan, Yu Gong, Philip S. Yu

Takin-VC: Expressive Zero-Shot Voice Conversion via Adaptive Hybrid Content Encoding and Enhanced Timbre Modeling (2025.acl-long)

Challenge:	Existing methods for short product title generation only consider textual information from long titles . MM-GAN incorporates image information and attribute tags from product, as well as textual info from original long titles.
Approach:	They propose a multi-modal generative adversarial network for short product title generation in E-commerce . they incorporate image information and attribute tags from product, as well as textual information from original long titles .
Outcome:	The proposed model outperforms state-of-the-art methods on a large-scale E-commerce dataset.

Copied to clipboard

Yang Yuguang, Yu Pan, Jixun Yao, Xiang Zhang, Jianhao Ye, Hongbin Zhou, Lei Xie, Lei Ma, Jianjun Zhao

LESA: Learnable LLM Layer Scaling-Up (2025.acl-long)

Challenge:	Expressive zero-shot voice conversion (VC) aims to modify source timbre to match unseen speaker . existing zero- shot VC systems struggle to reproduce paralinguistic information in highly expressive speech .
Approach:	They propose a framework for expressive zero-shot voice conversion that uses hybrid content encoding and memory-augmented context-aware timbre modeling.
Outcome:	The proposed framework surpasses state-of-the-art VC systems in speech naturalness, speaker similarity, and speaker similarness.

Copied to clipboard

Yifei Yang, Zouying Cao, Xinbei Ma, Yao Yao, Zhi Chen, Libo Qin, Hai Zhao

MessToClean: Evidence-Grounded Structure-Preserving Reconstruction for Real-World Degraded Exam Paper Images (2026.acl-long)

Challenge:	Existing methods for depth scaling-up rely on empirical heuristic rules for layer duplication, resulting in poor initialization and slower convergence during continual pre-training.
Approach:	They propose a method for learning latent parameters between layers by concatenating parameters from each layer and applying Singular Value Decomposition.
Outcome:	Experiments show that LESA outperforms baseline models with less than half the cost of existing methods.

Copied to clipboard

Jiayi Tuo, Cheng Tang, Zihan Wang, Chenyue Zhou, Yao Li, Yanbiao Ma, Chao Wang, Wei Dai, Mingxuan Wang, Shitong Qin, Ziwei Zhao

Prediction and Calibration: Complex Reasoning over Knowledge Graph with Bi-directional Directed Acyclic Graph Neural Network (2023.findings-acl)

Challenge:	Existing Multimodal Large Language Models (MLLMs) fail under RDEI, leading to disrupted structure and evidence-unsupported hallucinations.
Approach:	They propose a backbone-agnostic, evidence-driven pipeline that treats off-the-shelf MLLMs as interchangeable components to improve stem consistency and figure consistency.
Outcome:	The proposed pipeline improves stem consistency by 1.01-3.18%, figure consistency by 0.50-49.16%, and refusal F1 by 1.06-10.88% across question types.

Copied to clipboard

Yao Xu, Shizhu He, Li Cai, Kang Liu, Jun Zhao

Structured Data in NLP

Challenge:	Knowledge graphs (KGs) organize world knowledge as interlinked triples which describe entities and their relationships.
Approach:	They propose a bi-directional Directed Acyclic Graph neural network that splits the reasoning process into prediction and calibration.
Outcome:	The proposed model outperforms previous QE models on FB15k, FB16k-237, and NELL995 on prediction and calibration.

Apertus: Democratizing Open and Compliant LLMs for Global Language Environments (2026.acl-long)

Copied to clipboard

Alejandro Hernández-Cano, Alexander Hägele, Allen Hao Huang, Angelika Romanou, Antoni-Joan Solergibert, Barna Pásztor, Bettina Messmer, Dhia Garbaya, Eduard Frank Ďurech, Ido Hakimi, Juan Garcia Giraldo, Mete Ismayilzada, Negar Foroutan, Skander Moalla, Tiancheng Chen, Vinko Sabolčec, Yixuan Xu, Michael Aerni, Badr AlKhamissi, Inés Altemir Marinas, Mohammad Hossein Amani, Matin Ansaripour, Ilia Badanin, Harold Benoit, Emanuela Boros, Nicholas John Browning, Fabian Bösch, Maximilian Böther, Niklas Canova, Camille Challier, Clément Charmillot, Jonathan Coles, Jan Milan Deriu, Arnout Devos, Lukas Drescher, Daniil Dzenhaliou, Maud Ehrmann, Dongyang Fan, Simin Fan, Silin Gao, Miguel Gila, María Grandury, Diba Hashemi, Alexander Miserlis Hoyle, Jiaming Jiang, Mark Klein, Andrei Kucharavy, Anastasiia Kucherenko, Frederike Lübeck, Roman Machacek, Theofilos Ioannis Manitaras, Andreas Marfurt, Kyle Matoba, Simon Matrenok, Henrique Mendonça, Fawzi Roberto Mohamed, Syrielle Montariol, Luca Mouchel, Sven Najem-Meyer, Jingwei Ni, Gennaro Oliva, Matteo Pagliardini, Elia Palme, Andrei Panferov, Léo Paoletti, Marco Passerini, Ivan Pavlov, Auguste Poiroux, Kaustubh Ponkshe, Nathan Ranchin, Javier Rando, Mathieu Sauser, Jakhongir Saydaliev, Mukhammadali Sayfiddinov, Marian Schneider, Stefano Schuppli, Marco Scialanga, Andrei Semenov, Kumar Shridhar, Raghav Singhal, Anna Sotnikova, Alexander Sternfeld, Ayush Kumar Tarun, Paul Teiletche, Jannis Vamvas, Xiaozhe Yao, Hao Zhao, Alexander Ilic, Ana Klimovic, Andreas Krause, Caglar Gulcehre, David Rosenthal, Elliott Ash, Florian Tramèr, Joost VandeVondele, Livio Veraldi, Martin Rajman, Thomas C. Schulthess, Torsten Hoefler, Antoine Bosselut, Martin Jaggi, Imanol Schlag

Multilinguality

Challenge:	Apertus is a fully open suite of large language models (LLMs) designed to address responsibility shortcomings in today’s open model ecosystem, namely data responsibility and global representation.
Approach:	They propose to release a fully open suite of large language models (LLMs) that address data responsibility and global representation shortcomings in today’s open model ecosystem.
Outcome:	The proposed model is pretrained on openly available data and suppresses verbatim recall of data while retaining task performance.

Caution for the Environment: Multimodal LLM Agents are Susceptible to Environmental Distractions (2025.acl-long)

Copied to clipboard

Xinbei Ma, Yiting Wang, Yao Yao, Tongxin Yuan, Aston Zhang, Zhuosheng Zhang, Hai Zhao

Making RALM Robust to Irrelevant Contexts via Layer Knowledge Guided Attention (2025.findings-acl)

Challenge:	Experimental results show that multimodal GUI agents are susceptible to environmental distractions.
Approach:	They propose a scenario where both user and agent are benign and environment is not malicious . they implement an adversarial environment injection and analyze the approach to improve faithfulness .
Outcome:	The proposed approach improves faithfulness of multimodal large language model agents in a graphical user interface environment.

Copied to clipboard

Weijie Shi, Hao Chen, Jiaming Li, Yao Zhao, Yazhong Zhang, Qijin Chen, Jipeng Zhang, Ruiyuan Zhang, Jia Zhu, Jiajie Xu, Xiaofang Zhou

Self-supervised Product Title Rewrite for Product Listing Ads (2022.naacl-industry)

Challenge:	Large language models (LLMs) face factual hallucination and knowledge obsolescence when tackling knowledge-intensive tasks.
Approach:	They propose a layer-knowledge guided attention method which harnesses the layer-wise knowledge of large language models to optimize per-layer attention on useful passages.
Outcome:	The proposed method outperforms existing methods on RALM benchmarks.

Copied to clipboard

Xue Zhao, Dayiheng Liu, Junwei Ding, Liang Yao, Mahone Yan, Huibo Wang, Wenqing Yao

Towards Reverse Engineering of Language Models: A Survey (2025.findings-emnlp)

Challenge:	Existing work has investigated the title optimization for Product Listing Ads (PLAs) however, little work has examined the effectiveness of this method.
Approach:	They propose a method to rewrite product listing ads titles without considering the fluency and information priority.
Outcome:	The proposed solution reduces the cost and improves CTR in the offline test and real-world online test by a large amount.

Copied to clipboard

Xinpeng Ti, Wentao Ye, Zhifang Zhang, Junbo Zhao, Chang Yao, Lei Feng, Haobo Wang

GoT: Effective Graph-of-Thought Reasoning in Language Models (2024.findings-naacl)

Challenge:	Due to the vast amounts of data and computational resources required for model development, protecting the model’s parameters and training data has become an urgent and crucial concern.
Approach:	They define "reverse engineering" techniques as attacks on large language models and provide an in-depth analysis of them.
Outcome:	The proposed attacks are described as “reverse engineering” techniques on LMs and provide an introduction to existing protective strategies.

Copied to clipboard

Yao Yao, Zuchao Li, Hai Zhao

GoT-R1: Internalizing Graph-of-Thought via Structural Reinforcement for High-Density Reasoning (2026.findings-acl)

Challenge:	Recent advances in Large Language Models (LLMs) have been advancing at an unprecedented pace.
Approach:	They propose a graph-based approach which models human thought processes as a chain and as 'graphs' by representing thought units as nodes and connections between them as edges, they capture the non-sequential nature of human thinking and allows for a more realistic modeling of thought processes.
Outcome:	The proposed model improves on a text-only reasoning task and a multimodal reasoning task.

Copied to clipboard

Zuchao Li, Qiwei Li, Yao Yao, Hai Zhao, Lefei Zhang, Bo Du

Issues with Entailment-based Zero-shot Text Classification (2021.acl-short)

Challenge:	Chain-of-Thought reasoning suffers from an inherent mechanism flaw: linearity induces overthinking . emergence of Large Language Models (LLMs) has fundamentally redefined artificial intelligence .
Approach:	They propose a framework that replaces verbose linear trajectories with high-density reasoning graphs.
Outcome:	The proposed framework outperforms state-of-the-art models with reduced token overhead.

Copied to clipboard

Tingting Ma, Jin-Ge Yao, Chin-Yew Lin, Tiejun Zhao

Information Extraction & Text Mining

Challenge:	Pre-trained BERT models with no fine-tuning can yield competitive performance against BERT fine- tuned for NLI.
Approach:	They propose to use any target label into a sentence of hypothesis and verify whether it could be entailed by the input.
Outcome:	The proposed models perform better than models fine-tuned for BERT, but the results are in general negative.

Search-in-Context: Efficient Multi-Hop QA over Long Contexts via Monte Carlo Tree Search with Dynamic KV Retrieval (2025.findings-acl)

Copied to clipboard

Jiabei Chen, Guang Liu, Shizhu He, Kun Luo, Yao Xu, Jun Zhao, Kang Liu

NL2Formula: Generating Spreadsheet Formulas from Natural Language Queries (2024.findings-eacl)

Challenge:	Existing approaches to multihop question answering (MHQA) over long contexts are often neglecting explicit reasoning or incurring expensive computational costs due to full-attention mechanisms over long contextuals.
Approach:	They propose a framework that integrates Monte Carlo Tree Search (MCTS) with dynamic key-value retrieval to enable iterative, context-aware reasoning.
Outcome:	The proposed framework integrates Monte Carlo Tree Search (MCTS) with dynamic key-value (KV) retrieval to enable iterative, context-aware reasoning.

Copied to clipboard

Wei Zhao, Zhitao Hou, Siyuan Wu, Yan Gao, Haoyu Dong, Yao Wan, Hongyu Zhang, Yulei Sui, Haidong Zhang

PAR: Training-Free Positional Perturbation and Attention Recycling for Faithful OCR (2026.acl-long)

Challenge:	Creating spreadsheet formulas remains a tedious and error-prone task for many end-users . a novel task is proposed to generate spreadsheet formulae from a user's NL query .
Approach:	They propose a task to generate formulas that are grounded on a spreadsheet table given a Natural Language query as input.
Outcome:	The proposed task generates formulas that are grounded on a spreadsheet table, given a natural language query as input.

Copied to clipboard

Yao Yao, Manwen Liao, Weitian Zhang, Zuchao Li, Hai Zhao