Papers by Kan Li
LLM-Powered Benchmark Factory: Reliable, Generic, and Efficient (2026.acl-long)
Copied to clipboard
Peiwen Yuan, Shaoxiong Feng, Yiwei Li, Xinglin Wang, Yueqi Zhang, Jiayi Shi, Chuyi Tan, Boyuan Pan, Yao Hu, Kan Li
| Challenge: | Using generic and efficient benchmark generators, human annotators are limited by inefficiency . current benchmark generator methods rely on seed signals, leading to long cycles and high costs . |
| Approach: | They propose a framework to evaluate LLMs as generic benchmark generators and integrate them as BenchMaker. |
| Outcome: | The proposed framework achieves comparable performance to human-annotated benchmarks on most metrics. |
V-DPO: Mitigating Hallucination in Large Vision Language Models via Vision-Guided Direct Preference Optimization (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing large vision-language models suffer from hallucination due to over-reliance on the Large Language Model (LLM) backbone. |
| Approach: | They propose a method to improve visual context learning by using a large-scale preference learning algorithm to improve hallucination. |
| Outcome: | The proposed method improves on human-annotated hallucination datasets. |
Make Every Penny Count: Difficulty-Adaptive Self-Consistency for Cost-Efficient Reasoning (2025.findings-naacl)
Copied to clipboard
Xinglin Wang, Shaoxiong Feng, Yiwei Li, Peiwen Yuan, Yueqi Zhang, Chuyi Tan, Boyuan Pan, Yao Hu, Kan Li
| Challenge: | Existing decoding strategies for chain-of-thought reasoning do not exploit prior information about question difficulty. |
| Approach: | They propose a decoding strategy called self-consistency to improve reasoning performance by adjusting the number of samples based on the posterior distribution of a set of pre-samples. |
| Outcome: | The proposed method outperforms baseline methods on arithmetic, commonsense and symbolic reasoning tasks while achieving comparable performance. |
What’s Left Unsaid? Detecting and Correcting Misleading Omissions in Multimodal News Previews (2026.acl-long)
Copied to clipboard
| Challenge: | Existing efforts to detect factually incorrect content are omitted by creators who subtly reshape impressions by omitting crucial background context. |
| Approach: | They propose a multi-stage pipeline that simulates preview-based and context-based understanding and a OMGuard pipeline that combines interpretation-aware fine-tuning and rationale-guided misleading content correction. |
| Outcome: | The proposed framework lifts an 8B model’s detection accuracy to the level of a 235B LVLM while delivering stronger end-to-end correction. |
Focused Large Language Models are Stable Many-Shot Learners (2024.emnlp-main)
Copied to clipboard
Peiwen Yuan, Shaoxiong Feng, Yiwei Li, Xinglin Wang, Yueqi Zhang, Chuyi Tan, Boyuan Pan, Heda Wang, Yao Hu, Kan Li
| Challenge: | In-Context Learning (ICL) enables large language models to achieve rapid task adaptation by learning from demonstrations. |
| Approach: | They propose a training-free method that disperses model attention from the query . they propose 'focus' search strategy that uses model perplexity to ensure sufficient attention . |
| Outcome: | The proposed method achieves an average performance improvement of 5.2% over vanilla ICL and scales well with many-shot demonstrations. |
InsBank: Evolving Instruction Subset for Ongoing Alignment (2025.findings-emnlp)
Copied to clipboard
Jiayi Shi, Yiwei Li, Shaoxiong Feng, Peiwen Yuan, Xinglin Wang, Yueqi Zhang, Chuyi Tan, Boyuan Pan, Huan Ren, Yao Hu, Kan Li
| Challenge: | Recent studies emphasize that quality and diversity of instruction data are more crucial than quantity, highlighting the need to select diverse, high-quality subsets to reduce training costs. |
| Approach: | They propose to use a continuously updated repository to integrate the latest valuable instruction data with a progressive evolution framework to evolve InsBank over time. |
| Outcome: | The proposed framework outperforms baselines in InsBank evolution and extracts budget-specific subsets. |
Exploring Conditional Variational Mechanism to Pinyin Input Method for Addressing One-to-Many Mappings in Low-Resource Scenarios (2024.acl-short)
Copied to clipboard
| Challenge: | Experimental results demonstrate the superior performance of our method. |
| Approach: | They propose to leverage conditional variational mechanism to simplify pinyin IME . they employ a strategy that facilitates interaction between pinyan and Chinese character information . |
| Outcome: | The proposed method improves the performance of pinyin input method engine (IME) under low-resource conditions. |
Integrate the Essence and Eliminate the Dross: Fine-Grained Self-Consistency for Free-Form Language Generation (2024.acl-long)
Copied to clipboard
| Challenge: | Existing methods to improve output quality without aggregating input tokens are limited by the complexity of aggregation of responses. |
| Approach: | They propose to extract and integrate segment-level commonalities from candidate samples to enhance performance of LLMs in open-ended and reasoning tasks. |
| Outcome: | The proposed method improves performance on reasoning, code generation and mathematical reasoning tasks without requiring additional models and overlooking the knowledge present among the candidates. |
Modeling Complex Dialogue Mappings via Sentence Semantic Segmentation Guided Conditional Variational Auto-Encoder (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Existing efforts to identify and avoid CDM to facilitate dialogue learning failed to solve the problem. |
| Approach: | They propose a Sentence Semantic Segmentation guided Conditional Variational Auto-Encoder which can model and take advantage of the CDM data. |
| Outcome: | The proposed method can model and take advantages of the CDM data. |
Towards IP Intelligence: Benchmarking Large Language Models on Intellectual Property Knowledge and Practice (2026.findings-acl)
Copied to clipboard
Qiyao Wang, Guhong Chen, Hongbo Wang, Huaren Liu, Minghui Zhu, Zhifei Qin, Li Linwei, Yilin Yue, Shiqiang Wang, Jiayan Li, Wu Yihang, Ziqiang Liu, Longze Chen, Run Luo, Liyang Fan, Jiaming Li, Lei Zhang, Kan Xu, Hamid Alinejad-Rokny, Chengming Li, Shiwen Ni, Yuan Lin, Min Yang
| Challenge: | Existing datasets and benchmarks focus only on patents or cover limited aspects of the IP field, lacking alignment with real-world scenarios. |
| Approach: | They propose a bilingual IP task taxonomy and a large-scale bilingual benchmark to evaluate LLMs in real-world IP practice. |
| Outcome: | The proposed model achieves only 75.8% accuracy, indicating room for improvement . open-source IP and law-oriented models lag behind closed-source general-purpose models . |
Beyond One-Size-Fits-All: Tailored Benchmarks for Efficient Evaluation (2025.acl-long)
Copied to clipboard
Peiwen Yuan, Yueqi Zhang, Shaoxiong Feng, Yiwei Li, Xinglin Wang, Jiayi Shi, Chuyi Tan, Boyuan Pan, Yao Hu, Kan Li
| Challenge: | Existing efficient methods estimate performance of models on large benchmarks, but these methods rely on the assumption that target models have high prediction consistency with source models. |
| Approach: | They propose a method that conducts customized evaluation tailored to each target model. |
| Outcome: | The proposed method reduces the MAE of estimates by 31.4% on benchmarks across 300 models. |
Two-stage Generative Question Answering on Temporal Knowledge Graph Using Large Language Models (2024.findings-acl)
Copied to clipboard
| Challenge: | Temporal knowledge graph question answering (TKGQA) is one of the most challenging QA tasks due to the temporal constraints hidden in questions and the answers sought from dynamic structured knowledge. |
| Approach: | They propose a generative temporal knowledge graph question answering framework which guides LLMs to answer temporal questions through two phases: Subgraph Retrieval and Answer Generation. |
| Outcome: | The proposed framework exploits LLM’s intrinsic knowledge to mine temporal constraints and structural links in the questions without extra training, thus narrowing down the subgraph search space in both temporal and structural dimensions. |
DnA-Eval: Enhancing Large Language Model Evaluation through Decomposition and Aggregation (2025.coling-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) are scalable and economical evaluators, but how reliable they are is still under-explored. |
| Approach: | They propose a framework which breaks down the evaluation process into decomposition and aggregation stages based on pedagogical practices and provides an interpretable window for how well LLMs evaluate . |
| Outcome: | The proposed framework improves performance on a variety of meta-evaluation benchmarks by providing an interpretable window for how well LLMs evaluate . |
Emancipating Event Extraction from the Constraints of Long-Tailed Distribution Data Utilizing Large Language Models (2024.lrec-main)
Copied to clipboard
| Challenge: | Existing methods for EE depend on manual annotations, which are expensive and scarce. |
| Approach: | They propose to transform the event extraction task into multi-turn dialogues and a novel method for generating high-quality data. |
| Outcome: | The proposed methods significantly improve existing models’ performance with various paradigms and structures, especially on tail types. |
ECHo: A Visio-Linguistic Dataset for Event Causality Inference via Human-Centric Reasoning (2023.findings-emnlp)
Copied to clipboard
| Challenge: | ECHo is a diagnostic dataset of event causality inference grounded in visio-linguistic social scenarios. |
| Approach: | They propose a diagnostic dataset of event causality inference grounded in visio-linguistic social scenarios. |
| Outcome: | The proposed framework examines the reasoning capability of current AI systems on three human-centric tasks. |
Towards Fewer Hallucinations in Knowledge-Grounded Dialogue Generation via Augmentative and Contrastive Knowledge-Dialogue (2023.acl-short)
Copied to clipboard
| Challenge: | Existing knowledge-grounded dialogue generation models face the hallucination problem . Existing models generate inappropriate knowledge and generate inconsistent responses . |
| Approach: | They propose an Augmentative and Contrastive Knowledge Dialogue Expansion Framework to enhance existing knowledge dialogue models by polarizing optimization objectives and weak knowledge generation ability. |
| Outcome: | The proposed framework expands existing training sets and smooths the optimization objective that enables models to generate ground-truth with or without gold knowledge. |
Navigating Ethical Challenges in NLP: Hands-on strategies for students and researchers (2025.acl-tutorials)
Copied to clipboard
Luciana Benotti, Fanny Ducel, Karën Fort, Guido Ivetta, Zhijing Jin, Min-Yen Kan, Seunghun J. Lee, Minzhi Li, Margot Mieskes, Adriana Pagano
| Challenge: | This tutorial will equip participants with basic guidelines for thinking deeply about ethical issues . participants will gain practical experience on when to flag a paper for ethics review . |
| Approach: | This tutorial will equip participants with basic guidelines for thinking deeply about ethical issues . participants will gain practical experience on when to flag a paper for ethics review . |
| Outcome: | This tutorial will equip participants with basic guidelines for thinking deeply about ethical issues . participants will gain practical experience on when to flag a paper for ethics review . |
MLCopilot: Unleashing the Power of Large Language Models in Solving Machine Learning Tasks (2024.eacl-long)
Copied to clipboard
| Challenge: | Existing approaches to automating ML are time-consuming and difficult to understand for human developers. |
| Approach: | They propose a framework that leverages large language models to develop ML solutions for novel tasks. |
| Outcome: | The proposed framework bridges the gap between machine intelligence and human knowledge by exploiting state-of-the-art large language models. |
MVP-Bench: Can Large Vision-Language Models Conduct Multi-level Visual Perception Like Humans? (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing LVLMs perform visual perception at multiple levels, but they are not able to perform multi-level tasks. |
| Approach: | They propose a visual–language benchmark to evaluate LVLMs' perceptions . they use manipulated images to examine how LVLs can perform multi-level tasks . |
| Outcome: | The proposed model performs poorly on high-level perception tasks, the authors show . they also show that current models do not generalize in understanding semantics of synthetic images . |
Hierarchical Inductive Transfer for Continual Dialogue Learning (2022.findings-acl)
Copied to clipboard
| Challenge: | Existing frameworks for learning and deployment of neural dialogue models have been used for online chit-chat scenarios. |
| Approach: | They propose a hierarchical inductive transfer framework to learn and deploy dialogue skills continually and efficiently. |
| Outcome: | The proposed framework achieves comparable performance under deployment-friendly model capacity. |
Diversifying Neural Dialogue Generation via Negative Distillation (2022.naacl-main)
Copied to clipboard
| Challenge: | Existing approaches to generate generic responses are ignoring low-frequency but generic responses and bringing low- frequency but meaningless responses. |
| Approach: | They propose a negative training paradigm that reminds dialogue models not to generate high-frequency responses during training. |
| Outcome: | The proposed method outperforms previous methods in the generic response problem while minimizing low-frequency but meaningless responses. |
CogLM: Tracking Cognitive Development of Large Language Models (2025.naacl-long)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have recently shown remarkable abilities across a wide variety of tasks, but few studies have explored the reasons behind the evolutionary relationship among various abilities. |
| Approach: | They construct a benchmark CogLM based on Piaget's Theory of Cognitive Development (PTC) which measures the cognitive levels of Large Language Models (LLMs) using 1,220 questions spanning 10 cognitive abilities crafted by more than 20 human experts. |
| Outcome: | The proposed framework provides a comprehensive testbed for the cognitive levels of LLMs. |
Molweni: A Challenge Multiparty Dialogues-based Machine Reading Comprehension Dataset with Discourse Structure (2020.coling-main)
Copied to clipboard
| Challenge: | Multiparty dialog applications such as discourse parsing and meeting summarization are now mainstream research. |
| Approach: | They propose to annotate a machine reading comprehension dataset with discourse structure built over multiparty dialog using a modified Segmented Discourse Representation Theory (SDRT) style. |
| Outcome: | The proposed dataset contributes large-scale discourse dependency annotations in a modified Segmented Discourse Representation Theory (SDRT) style for all of its multiparty dialogs, and achieves only 67.7% F1 on Molweni’s questions, a 20+% significant drop as compared against its SQuAD 2.0 performance. |
ChatCRS: Incorporating External Knowledge and Goal Guidance for LLM-based Conversational Recommender Systems (2025.findings-naacl)
Copied to clipboard
| Challenge: | Experimental results show that ChatCRS improves language quality and informativeness by 17% and proactivity by 27%. |
| Approach: | They propose a framework to decompose the CRS task into several sub-tasks . they propose 'knowledge retrieval agent' and 'goal-planning agent' |
| Outcome: | The proposed framework improves language quality and informativeness by 17% and proactivity by 27% on two multi-goal CRS datasets. |
Speculative Decoding for Multi-Sample Inference (2025.findings-emnlp)
Copied to clipboard
Yiwei Li, Jiayi Shi, Shaoxiong Feng, Peiwen Yuan, Xinglin Wang, Yueqi Zhang, Ji Zhang, Chuyi Tan, Boyuan Pan, Yao Hu, Kan Li
| Challenge: | Speculative decoding method exploits consensus of parallel reasoning paths to synthesize high-quality draft tokens without auxiliary models or external databases. |
| Approach: | They propose a speculative decoding method that exploits the consensus of parallel reasoning paths to synthesize high-quality draft tokens without auxiliary models or external databases. |
| Outcome: | The proposed method exploits the intrinsic consensus of parallel reasoning paths to synthesize high-quality draft tokens without auxiliary models or databases. |
Generating Relevant and Coherent Dialogue Responses using Self-Separated Conditional Variational AutoEncoders (2021.acl-long)
Copied to clipboard
| Challenge: | Conditional Variational AutoEncoders (CVAE) can enhance the diversity and informativeness of responses in open-domain dialogue generation tasks. |
| Approach: | They propose a Conditional Variational AutoEncoder (CVAE) that regularizes latent variables and introduces group information to regularize them. |
| Outcome: | Empirical results show that the proposed model can significantly boost responses in well-established open-domain dialogue datasets. |
Learning to Select In-Context Demonstration Preferred by Large Language Model (2025.findings-acl)
Copied to clipboard
| Challenge: | In-context learning (ICL) enables large language models to perform tasks with only a few examples as demonstrations. |
| Approach: | They propose a generative preference learning framework that leverages LLM feedback to directly optimize demonstration selection for ICL. |
| Outcome: | Experiments on 19 datasets across 11 task categories show that GenICL achieves superior performance than existing methods in selecting the most effective demonstrations. |
Regularizing Dialogue Generation by Imitating Implicit Scenarios (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing models for dialogue generation lack the flexibility to handle such freedoms. |
| Approach: | They propose to take into account dialogue history and future conversation to implicitly reconstruct the scenario knowledge. |
| Outcome: | The proposed approach outperforms state-of-the-art models on diversity and relevance and expresses scenario-specific knowledge. |
EASYTOOL: Enhancing LLM-based Agents with Concise Tool Instruction (2025.naacl-long)
Copied to clipboard
| Challenge: | EASYTOOL combines tools from diverse tool documentation into a single tool instruction. |
| Approach: | They propose a framework that transforms tool documentation into a unified tool instruction. |
| Outcome: | EASYTOOL combines extensive tool documentation into a concise tool instruction . it reduces token consumption and improves performance of LLM-based agents . |
Exploring Pre-trained Language Models for Event Extraction and Generation (P19-1)
Copied to clipboard
| Challenge: | Existing methods to extract event data are laborious to create and limited in size. |
| Approach: | They propose an event extraction model to overcome the roles overlap problem by separating the argument prediction in terms of roles. |
| Outcome: | The proposed method surpasses existing methods on the ACE2005 dataset and improves on the previous methods. |
From Sub-Ability Diagnosis to Human-Aligned Generation: Bridging the Gap for Text Length Control via MarkerGen (2025.acl-long)
Copied to clipboard
Peiwen Yuan, Chuyi Tan, Shaoxiong Feng, Yiwei Li, Xinglin Wang, Yueqi Zhang, Jiayi Shi, Boyuan Pan, Yao Hu, Kan Li
| Challenge: | Existing methods to control text length are lacking in LCTG, posing a major limitation for practical applications. |
| Approach: | They propose a plug-and-play approach that decomposes LCTG sub-abilities with human patterns as reference and performs detailed error analysis. |
| Outcome: | The proposed method significantly improves LCTG across various settings, exhibiting outstanding effectiveness and generalizability. |
Learning Joint Structural and Temporal Contextualized Knowledge Embeddings for Temporal Knowledge Graph Completion (2023.findings-acl)
Copied to clipboard
| Challenge: | Existing methods that incorporate time information into static knowledge graph embedding ignore the contextual nature of the TKG structure. |
| Approach: | They propose a method that employs pre-trained language models to learn joint Structural and Temporal Contextualized Knowledge Embeddings. |
| Outcome: | The proposed method is superior to existing methods that ignore the contextual nature of the TKG structure. |
Revisiting Self-Consistency from Dynamic Distributional Alignment Perspective on Answer Aggregation (2025.findings-acl)
Copied to clipboard
Yiwei Li, Ji Zhang, Shaoxiong Feng, Peiwen Yuan, Xinglin Wang, Jiayi Shi, Yueqi Zhang, Chuyi Tan, Boyuan Pan, Yao Hu, Kan Li
| Challenge: | Existing studies on self-consistency show that it improves reasoning abilities by aggregating diverse stochastic samples. |
| Approach: | They propose a confidence-driven mechanism that dynamically calibrates temperature to align with high probability modes. |
| Outcome: | The proposed method outperforms fixed-diversity baselines on reasoning tasks and improves both average and best-case performance. |
CoAnnotating: Uncertainty-Guided Work Allocation between Human and Large Language Models for Data Annotation (2023.emnlp-main)
Copied to clipboard
| Challenge: | Annotated data plays a critical role in training models and evaluating their performance. |
| Approach: | They propose a paradigm for Human-LLM co-annotation of unstructured texts at scale that utilizes uncertainty to estimate LLMs’ annotation capability. |
| Outcome: | The proposed model outperforms existing models on many text-annotation tasks with up to 21% performance improvement over random baseline. |
Large Language Models Are Still Misled by Simple Bias Ensembles (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing benchmarks for large language models are constrained to datasets where each sample is manually injected with only one type of bias. |
| Approach: | They propose a multi-bias benchmark where each sample contains multiple types of biases. |
| Outcome: | The proposed benchmark shows that existing LLMs and debiasing methods perform poorly on this benchmark, highlighting the challenge of eliminating compounded biases. |
Beyond the Crowd: LLM-Augmented Community Notes for Governing Health Misinformation (2026.acl-long)
Copied to clipboard
| Challenge: | X (formerly Twitter) users can flag misleading posts, attach contextual notes, and rate the notes’ helpfulness, but there is a significant latency in Community Notes, which is unable to provide accurate notes. |
| Approach: | They propose a framework that augments Community Notes for faster and more reliable health misinformation governance. |
| Outcome: | The proposed framework outperforms human contributors in correctness, helpfulness, and evidence utility in health misinformation surges. |
GL-CLeF: A Global–Local Contrastive Learning Framework for Cross-lingual Spoken Language Understanding (2022.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to zero-shot cross-lingual spoken language understanding rely on shared parameters, which can only perform implicit alignment across languages. |
| Approach: | They propose a global-local contrastive learning framework to achieve a fine-grained cross-lingual transfer . they employ bilingual dictionaries to construct multilingual views of the same utterance . |
| Outcome: | Experiments on MultiATIS++ show that GL-CLeF achieves the best performance . GL is based on dictionaries and encourages representations to be more similar than negative example pairs . |
UNO-DST: Leveraging Unlabelled Data in Zero-Shot Dialogue State Tracking (2024.findings-naacl)
Copied to clipboard
| Challenge: | Existing methods for zero-shot dialogue state tracking (DST) ignore unlabelled data in the target domain. |
| Approach: | They propose to transform zero-shot dialogue state tracking into few-shot DST by utilising unlabelled data via joint and self-training methods. |
| Outcome: | The proposed method improves joint goal accuracy by 8% on general language models in zero-shot scenarios, and can be used in many domains. |
Generative Dense Retrieval: Memory Can Be a Burden (2024.eacl-long)
Copied to clipboard
| Challenge: | Empirical results show that Generative Dense Retrieval (GDR) achieves an average of 3.0 R@100 improvement on NQ dataset under multiple settings and has better scalability. |
| Approach: | They propose a Generative Dense Retrieval paradigm that auto-decodes document identifiers given a query and uses memory to avoid memory confusion. |
| Outcome: | Empirical results show that the proposed paradigm improves on the small-scale corpora and improves scalability. |
Poor-Supervised Evaluation for SuperLLM via Mutual Consistency (2024.findings-acl)
Copied to clipboard
| Challenge: | evaluating superLLMs is especially difficult because of their intelligence-intensive nature. |
| Approach: | They propose an evaluation benchmark with accurate labels for SuperLLMs whose capabilities surpass those of humans . they first prove that consistency between model under evaluation and reference model can equalize the true capabilities of the model to be evaluated . |
| Outcome: | The proposed evaluation benchmarks can assess the true capabilities of the model to be evaluated without accurate labels. |
Dynamic Stochastic Decoding Strategy for Open-Domain Dialogue Generation (2024.findings-acl)
Copied to clipboard
| Challenge: | Stochastic sampling strategies are not widely used in open-domain dialogue systems. |
| Approach: | They propose a dynamic decoding strategy which can adjust the decoding space w.r.t. different contexts. |
| Outcome: | The proposed decoding strategy can improve the performance of pre-trained models when coupled with four well-used stochastic decoding algorithms. |