Papers by Deng Cai

Towards Self-Evolving Agents: Enabling Autonomy through Interactive Experience Refinement (2026.findings-acl)

Challenge:	Abstract Meaning Representation (AMR) parsing is a broad-coverage semantic formalism that encodes the meaning of a sentence as a rooted, directed, and labeled graph.
Approach:	They propose to use existing English parser to learn and improve multilingual AMR parsers . their results show that noisy input and precise output are key to successful distillation .
Outcome:	The proposed model outperforms the current state-of-the-art English-only parser on four different languages.

Copied to clipboard

Cheng Yang, Xuemeng Yang, Licheng Wen, Daocheng Fu, Jianbiao Mei, Rong Wu, Pinlong Cai, Yufan Shen, Nianchen Deng, Jia Xu, Botian Shi, Yu Qiao, Haifeng Li

Challenge:	Large Language Models struggle with complex, multi-step operational tasks because they remain static during inference and cannot learn from past experience.
Approach:	They propose a framework that organizes cross-domain insights to facilitate orchestration of long-horizon workflows.
Outcome:	The proposed framework outperforms existing methods on the TAC productivity benchmark and shows strong cross-task transferability.

DentalGPT: Incentivizing Multimodal Reasoning in Dentistry (2026.findings-acl)

Copied to clipboard

Zhenyang Cai, Jiaming Zhang, Junjie Zhao, Ziyi Zeng, Yanchao Li, Liang Jingyi, Junying Chen, Yunjin Yang, Jiajun You, Shuzhi Deng, null Xieruiqiii, Yuanting Chen, Xiangyi Feng, Jianquan Li, Liangyi Chen, Junwen Wang, Shan Jiang, Benyou Wang

Consecutive Batch Model Editing with HooK Layers (2024.emnlp-main)

Challenge:	Current multimodal large language models (MLLMs) show limited understanding of dental images.
Approach:	They propose a dental-specialized multimodal large language model trained via staged multimodal alignment and reinforcement learning.
Outcome:	The proposed model outperforms state-of-the-art models on disease classification and dental VQA tasks.

Copied to clipboard

Shuaiyi Li, Yang Deng, Deng Cai, Hongyuan Lu, Liang Chen, Wai Lam

Challenge:	Existing models that retrain are time- and resource-consuming, but they lack the memory to support sequential and batch editing.
Approach:	They propose a model editing method that supports sequential and batch editing . they use a small amount of memory to store several hook layers that remain unchanged over time .
Outcome:	The proposed method is memory-friendly and can store hook layers that remain unchanged over time.

Retrieval-guided Dialogue Response Generation via a Matching-to-Generation Framework (D19-1)

Copied to clipboard

Deng Cai, Yan Wang, Wei Bi, Zhaopeng Tu, Xiaojiang Liu, Shuming Shi

Translating a Math Word Problem to a Expression Tree (D18-1)

Challenge:	generative models for end-to-end sequence generation have been shown promising for this task . however, how to precisely extract a skeleton and how to effectively train a retrieval-guided response generator is still challenging.
Approach:	They propose a framework where skeleton extraction is made by an interpretable matching model and a retrieval-guided response generator is followed by a separate generator.
Outcome:	The proposed framework outperforms baseline models in a variety of experiments.

Copied to clipboard

Lei Wang, Yan Wang, Deng Cai, Dongxiang Zhang, Xiaojiang Liu

A Thorough Examination of Decoding Methods in the Era of LLMs (2024.emnlp-main)

Challenge:	Sequence-to-sequence (SEQ2SEQ) models have been successfully applied to automatic math word problem solving.
Approach:	They propose an equation normalization method to normalize duplicated equations and propose an ensemble model to combine their advantages.
Outcome:	The proposed model outperforms the previous state-of-the-art models on the math word problem solving.

Copied to clipboard

Chufan Shi, Haoran Yang, Deng Cai, Zhisong Zhang, Yifan Wang, Yujiu Yang, Wai Lam

Reasons to Reject? Aligning Language Models with Judgments (2024.findings-acl)

Challenge:	Decoding methods are essential for converting language models from next-token predictors into practical task solvers.
Approach:	They propose to evaluate decoding methods in general-purpose large language models . they find that decoding method performance is notably task-dependent .
Outcome:	The proposed methods perform task-dependently and are influenced by alignment, model size, and quantization.

Copied to clipboard

Weiwen Xu, Deng Cai, Zhisong Zhang, Wai Lam, Shuming Shi

SpecEdit: A Spectral Approach for Multi-Round Knowledge Editing (2026.findings-acl)

Challenge:	a new framework for aligning large language models with judgments is proposed to help with alignment . a framework that allows for fine-grained inappropriate content detection and correction based on judgments . large language model alignment is critical for making artificial intelligence a reliable ally for humanity .
Approach:	They propose a framework that allows for fine-grained inappropriate content detection and correction based on judgments.
Outcome:	The proposed framework beats the 175B DaVinci003 and improves on AlpacaEval using judgments.

Copied to clipboard

Junxian Liu, Zikun Deng, Xin Wu, Dazhen Deng, Yi Cai

Explainability & Interpretability in NLP

Challenge:	Multi-round knowledge editing suffers from performance degradation as edits accumulate . intrinsic knowledge of model and historical edit memories are naively coupled during editing . SpecEdit improves model editing performance by reducing destructive coupling .
Approach:	They propose a spectral-based model editing module that integrates into existing editing methods without altering their original optimization procedures.
Outcome:	The proposed model improves performance on multiple LLMs and editing methods.

RuleEdit: Towards Rule-Level Knowledge Generalization to Mitigate Over-Editing in Large Language Models (2025.findings-acl)

Copied to clipboard

Bihan Zhou, HaoPeng Ren, Li Yuan, Yi Cai, Liuwen Cao, Zikun Deng

Light-R1: Curriculum SFT, DPO and RL for Long COT from Scratch and Beyond (2025.acl-industry)

Challenge:	Existing knowledge editing methods focus on instance-level editing, which is prone to knowledge degradation and general ability deterioration due to redundant instance-specific modifications.
Approach:	They propose a rule-level editing method that generalizes rule-derived knowledge to update rule-based instances.
Outcome:	The proposed method improves portability and performance over baselines for LLaMA-2-7B on RULEmix.

Copied to clipboard

Liang Wen, Yunke Cai, Fenrui Xiao, Xin He, Qi An, Zhenyu Duan, Yimin Du, Junchen Liu, Tanglifu Tanglifu, Xiaowei Lv, Haosheng Zou, Yongchao Deng, Shousheng Jia, Xiangzheng Zhang

CADReview: Automatically Reviewing CAD Programs with Error Detection and Correction (2025.acl-long)

Challenge:	Experimental results show that opensource curriculum training is more effective when distinct datasets are available for different training stages.
Approach:	They propose an opensource suite for training long reasoning models using publicdata and models.
Outcome:	The proposed model outperforms DeepSeek-R1-DistillQwen-32B models in math reasoning.

Copied to clipboard

Jiali Chen, Xusen Hei, HongFei Liu, Yuancheng Wei, Zikun Deng, Jiayuan Xie, Yi Cai, Li Qing

A Frustratingly Simple Decoding Method for Neural Text Generation (2024.lrec-main)

Challenge:	Computer-aided design (CAD) is crucial in prototyping 3D objects through geometric instructions.
Approach:	They propose a CAD review task to automatically detect and correct potential errors . they propose CAD program repairer framework to provide helpful feedback on error correction .
Outcome:	The proposed framework outperforms existing MLLMs in detecting errors and providing feedback on error correction.

Copied to clipboard

Haoran Yang, Deng Cai, Huayang Li, Wei Bi, Wai Lam, Shuming Shi

AMR Parsing via Graph-Sequence Iterative Inference (2020.acl-main)

Challenge:	Neural text generation is notorious for repetitive loops and tedious outputs.
Approach:	They propose a method that penalizes future generation of repetitive content . they construct an anti-LM based on previously generated text .
Outcome:	The proposed method outperforms established baselines in terms of generation quality, decoding speed, and universality.

Copied to clipboard

Deng Cai, Wai Lam

Self-Reasoning Language Models: Unfold Hidden Reasoning Chains with Few Reasoning Catalyst (2025.findings-acl)

Challenge:	Abstract Meaning Representation (AMR) parsing is a broad-coverage semantic formalism that encodes the meaning of a sentence as a rooted, directed, labeled graph.
Approach:	They propose a model that treats AMR parsing as a series of dual decisions on the input sequence and the incrementally constructed graph.
Outcome:	The proposed model outperforms existing models by large margins on both input sequence and output graph.

Copied to clipboard

Hongru Wang, Deng Cai, Wanjun Zhong, Shijue Huang, Jeff Z. Pan, Zeming Liu, Kam-Fai Wong

Discourse Marker Augmented Network with Reinforcement Learning for Natural Language Inference (P18-1)

Challenge:	Recent studies have demonstrated that inference-time scaling increases performance of Large Language Models (LLMs) in various reasoning tasks such as mathematics and complex question answering by increasing the length of Chain-of-Thought (CoT).
Approach:	They propose a model which synthesizes longer CoT data and iteratively improves performance through self-training by incorporating a few demonstration examples.
Outcome:	The proposed model achieves an average improvement of more than +2.5 points across five reasoning tasks: MMLU, GSM8K, ARC-C, HellaSwag, and BBH on two backbone models.

Copied to clipboard

Boyuan Pan, Yazheng Yang, Zhou Zhao, Yueting Zhuang, Deng Cai, Xiaofei He

TextBind: Multi-turn Interleaved Multimodal Instruction-following in the Wild (2024.findings-acl)

Challenge:	Existing approaches to natural language inference focus on interaction architectures of sentences . but, we propose to transfer knowledge from discourse markers to augment the model .
Approach:	They propose to transfer knowledge from discourse markers to augment the quality of the NLI model.
Outcome:	The proposed method achieves state-of-the-art performance on large-scale datasets.

Copied to clipboard

Huayang Li, Siheng Li, Deng Cai, Longyue Wang, Lemao Liu, Taro Watanabe, Yujiu Yang, Shuming Shi

Exploring Compositional Generalization of Large Language Models (2024.naacl-srw)

Challenge:	Large language models with instruction-following capabilities have revolutionized the field of artificial intelligence.
Approach:	They propose an annotation-free framework for empowering large language models with instruction-following capabilities.
Outcome:	The proposed framework generates multi-turn multimodal instruction-response conversations from a language model.

Copied to clipboard

Haoran Yang, Hongyuan Lu, Wai Lam, Deng Cai

The World is Not Binary: Learning to Rank with Grayscale Data for Dialogue Response Selection (2020.emnlp-main)

Challenge:	a recent study has found that large language models can generalize compositional instructions from simple instructions to complex ones.
Approach:	They study the generalization ability of large language models with respect to compositional instructions . they first construct a dataset with the help of ChatGPT guided by the self-instruct technique .
Outcome:	The proposed model can generalize from simple instructions to more intricate ones, the authors show . their results show that training LLMs on higher-order compositional instructions improves performance on lower-order ones, but not on higher order ones.

Copied to clipboard

Zibo Lin, Deng Cai, Yan Wang, Xiaojiang Liu, Haitao Zheng, Shuming Shi

Information Retrieval

Challenge:	Existing approaches to learning-to-rank response selection are suboptimal due to ignorance of diversity of response quality.
Approach:	They propose to use off-the-shelf response retrieval models as automatic grayscale data generators to train response selection models.
Outcome:	The proposed approach can be automated without human effort on grayscale data.

Skeleton-to-Response: Dialogue Generation Guided by Retrieval Memory (N19-1)

Copied to clipboard

Deng Cai, Yan Wang, Wei Bi, Zhaopeng Tu, Xiaojiang Liu, Wai Lam, Shuming Shi

InfiniteICL: Breaking the Limit of Context Window Size via Long Short-term Memory Transformation (2025.findings-acl)

Challenge:	Existing generative dialogue models generate responses from input queries . however, the results are limited and the models are unsatisfactory .
Approach:	They propose a framework which exploits retrieval results via a skeleton-to-response paradigm . they extract a query skelet and use it to generate a new skele and response .
Outcome:	The proposed approach significantly improves the informativeness of the generated responses.

Copied to clipboard

Bowen Cao, Deng Cai, Wai Lam

Multi-Task Pre-Training for Plug-and-Play Task-Oriented Dialogue System (2022.acl-long)

Challenge:	InfiniteICL is a framework that parallels context and parameters in large language models with short- and long-term memory in human cognitive systems.
Approach:	They propose a framework that parallels context and parameters in large language models with short- and long-term memory in human cognitive systems and enables infinite context integration.
Outcome:	The proposed framework reduces context length by 90% while achieving 103% average performance of full-context prompting across fact recall, grounded reasoning, and skill acquisition tasks.

Copied to clipboard

Yixuan Su, Lei Shu, Elman Mansimov, Arshit Gupta, Deng Cai, Yi-An Lai, Yi Zhang

Assessing Dialogue Systems with Distribution Distances (2021.findings-acl)

Challenge:	Existing pre-trained language models often form a cascaded generation problem . this can lead to error accumulation across different sub-tasks and greater data annotation overhead.
Approach:	They propose a plug-and-play model for task-oriented dialogue that learns primary TOD task completion skills from heterogeneous dialog corpora.
Outcome:	The proposed model learns primary TOD task completion skills from heterogeneous dialog corpora.

Copied to clipboard

Jiannan Xiang, Yahui Liu, Deng Cai, Huayang Li, Defu Lian, Lemao Liu

Natural Language Interfaces

Challenge:	Existing evaluation metrics focus on turnlevel quality, which is not well suited for open-end dialogue tasks.
Approach:	They propose to measure the performance of a dialogue system by computing the distributionwise distance between its generated conversations and real-world conversations.
Outcome:	The proposed metrics correlate better with human judgments than existing metrics on dialogue systems.

Neural Machine Translation with Monolingual Translation Memory (2021.acl-long)

Copied to clipboard

Deng Cai, Yan Wang, Huayang Li, Wai Lam, Lemao Liu

Multilinguality

Challenge:	Existing work has shown that Translation Memory (TM) can boost the performance of Neural Machine Translation (NMT)
Approach:	They propose a framework that uses monolingual memory and performs learnable memory retrieval in a cross-lingual manner.
Outcome:	The proposed framework outperforms strong TM-augmented NMT baselines using bilingual TM and outperformed existing models in low-resource and domain adaptation scenarios.

Can LLMs Act as Historians? Evaluating Historical Research Capabilities of LLMs via the Chinese Imperial Examination (2026.acl-long)

Copied to clipboard

Lirong Gao, Zeqing Wang, Yuyan Cai, Jiayi Deng, Yanmei Gu, Yiming Zhang, Jia Zhou, Yanfei Zhang, Junbo Zhao

Effidit: An Assistant for Improving Writing Efficiency (2023.acl-demo)

Challenge:	Existing benchmarks assess basic knowledge breadth or lexical understanding, failing to capture higher-order skills that are central to historical research.
Approach:	They propose a benchmark anchored in the Chinese Imperial Examination system that assesses historical knowledge and lexical understanding.
Outcome:	The new benchmark aims to assess the ability of LLMs to process historical materials and documents.

Copied to clipboard

Shuming Shi, Enbo Zhao, Wei Bi, Deng Cai, Leyang Cui, Xinting Huang, Haiyun Jiang, Duyu Tang, Kaiqiang Song, Longyue Wang, Chenyan Huang, Guoping Huang, Yan Wang, Piji Li

Green & Sustainable NLP

Challenge:	Effidit is a digital writing assistant that provides three modules to help users write faster and more efficiently.
Approach:	They present Effidit, a digital writing assistant that provides three modules to help users write higher-quality text more efficiently.
Outcome:	Effidit expands the capabilities of a typical writing assistant by providing three modules . Effit can help users create their own text faster and more efficiently .

WatME: Towards Lossless Watermarking Through Lexical Redundancy (2024.acl-long)

Copied to clipboard

Liang Chen, Yatao Bian, Yang Deng, Deng Cai, Shuaiyi Li, Peilin Zhao, Kam-Fai Wong

Challenge:	Existing methods for text watermarking rely on arbitrary vocabulary partitioning during decoding, which compromises the availability of suitable tokens and significantly degrades the quality of responses.
Approach:	They propose a method that leverages linguistic prior knowledge of lexical redundancies in LLM vocabularies to seamlessly integrate watermarks.
Outcome:	The proposed approach preserves the expressive power of large language models while preserving watermark detectability.

Retrofitting Multilingual Sentence Embeddings with Abstract Meaning Representation (2022.emnlp-main)

Copied to clipboard

Deng Cai, Xin Li, Jackie Chun-Sing Ho, Lidong Bing, Wai Lam

Exploiting Reasoning Chains for Multi-hop Science Question Answering (2021.findings-emnlp)

Challenge:	Existing studies on multilingual sentence embeddings focus on cross-lingual semantic textual similarity and transfer tasks.
Approach:	They propose a method to improve existing multilingual sentence embeddings with Abstract Meaning Representation (AMR) . they compare existing multi-lingual sentence embedded with AMR and improve their versions by reducing the surface variations across different languages and expressions.
Outcome:	The proposed method improves state-of-the-art multilingual sentence embeddings on transfer tasks and semantic textual similarity tests.

Copied to clipboard

Weiwen Xu, Yang Deng, Huihui Zhang, Deng Cai, Wai Lam

Responsible & Trustworthy NLP

Challenge:	Existing frameworks for multi-hop Science question answering do not require corpus-specific annotations.
Approach:	They propose a chain-guided retriever-reader framework that performs explainable reasoning without corpus annotations.
Outcome:	The proposed framework performs explainable reasoning without corpus-specific annotations . it is shown to be effective on OpenBookQA and ARC-Challenge .

Enhancing Chain-of-Thought Reasoning with Critical Representation Fine-tuning (2025.acl-long)

Copied to clipboard

Chenxi Huang, Shaotian Yan, Liang Xie, Binbin Lin, Sinan Fan, Yue Xin, Deng Cai, Chen Shen, Jieping Ye

Charge-Based Prison Term Prediction with Deep Gating Network (D19-1)

Challenge:	Representation Fine-tuning (ReFT) is a proposed method for improving parameter efficiency . however, it yields suboptimal performance, as fixed-position representations have uncertain impact on outputs .
Approach:	They propose a method that fine-tunes critical representations in a low-rank linear subspace while freezing the base model.
Outcome:	The proposed method improves accuracy of LLaMA-2-7B and ReFT by 18.2 and 3.8 on GSM8K.

Copied to clipboard

Huajie Chen, Deng Cai, Wei Dai, Zehui Dai, Yadong Ding

Challenge:	Existing work merely predicts the total prison term, but in reality a defendant is often charged with multiple crimes.
Approach:	They propose a charge-based prison term prediction task that better fits real needs and makes it more accurate and interpretable.
Outcome:	The proposed method achieves state-of-the-art performance for charge-specific feature selection and aggregation.

Cross-lingual Contextualized Phrase Retrieval (2024.findings-emnlp)

Copied to clipboard

Huayang Li, Deng Cai, Zhi Qu, Qu Cui, Hidetaka Kamigaito, Lemao Liu, Taro Watanabe

Multilinguality

Challenge:	Phrase-level dense retrieval has shown many appealing characteristics in downstream NLP tasks.
Approach:	They propose a task formulation of dense retrieval, cross-lingual contextualized phrase retrieval . they extract pairs of cross-linguistic phrases using word alignment information .
Outcome:	The proposed task formulation surpasses baselines on the phrase retrieval task and a downstream task, i.e., machine translation, and achieves top-1 accuracy 13 points higher.

Dynamic Semantic Graph Construction and Reasoning for Explainable Multi-hop Science Question Answering (2021.findings-acl)

Copied to clipboard

Weiwen Xu, Huihui Zhang, Deng Cai, Wai Lam

Core Semantic First: A Top-down Approach for AMR Parsing (D19-1)

Challenge:	Existing approaches suffer from low confidence when retrieving evidence facts to fill the knowledge gap and lack transparent reasoning process.
Approach:	They propose a framework to exploit more valid facts while obtaining explainability for multi-hop question answering at web scale by dynamically constructing a semantic graph and reasoning over it.
Outcome:	The proposed framework surpasses existing approaches while maintaining high explainability on OpenBookQA and ARC-Challenge.

Copied to clipboard

Deng Cai, Wai Lam

FlightGPT: Towards Generalizable and Interpretable UAV Vision-and-Language Navigation with Vision-Language Models (2025.emnlp-main)

Challenge:	Abstract Meaning Representation (AMR) parsing is a semantic formalism that encodes the meaning of a sentence as a rooted labeled directed graph.
Approach:	They propose a scheme for parsing text into its Abstract Meaning Representation (AMR) using Graph Spanning based Parsing.
Outcome:	The proposed scheme achieves state-of-the-art on the latest AMR sembank and no heuristic graph re-categorization is adopted.

Copied to clipboard

Hengxing Cai, Jinhan Dong, Jingjun Tan, Jingcheng Deng, Sihang Li, Zhifeng Gao, Haidong Wang, Zicheng Su, Agachai Sumalee, Renxin Zhong

Reinforced Dynamic Reasoning for Conversational Question Generation (P19-1)

Challenge:	Existing methods for vision-and-language navigation struggle with insufficient multimodal fusion, weak generalization, and poor interpretability.
Approach:	They propose a framework for UAV vision-and-language navigation that integrates natural language instructions with visual observations to improve multimodal fusion and interpretability.
Outcome:	The proposed framework achieves state-of-the-art performance across all scenarios, with a 9.22% higher success rate than the strongest baseline in unseen environments.

Copied to clipboard

Boyuan Pan, Hao Li, Ziyu Yao, Deng Cai, Huan Sun

Natural Language Interfaces

Challenge:	Empirical results on the recently released CoQA dataset demonstrate the effectiveness of our method . large-scale highquality conversational question answering datasets such as CoQA and QuAC can help train models to answer sequential questions.
Approach:	They propose a task called Conversational Question Generation which generates a question based on a passage and a conversation history to generate the next question.
Outcome:	The proposed method is based on a question-answering style conversation dataset . it can be used to generate meaningful questions on QA and SQuAD datasets .

Disperse-Then-Merge: Pushing the Limits of Instruction Tuning via Alignment Tax Reduction (2024.findings-acl)

Copied to clipboard

Tingchen Fu, Deng Cai, Lemao Liu, Shuming Shi, Rui Yan

Dialogue Response Selection with Hierarchical Curriculum Learning (2021.acl-long)

Challenge:	Pre-trained language models may not follow human instructions and produce toxic, hallucinated, or biased content.
Approach:	They propose a disperse-then-merge framework that dispersers instruction-following data into portions and trains multiple sub-models using different data portions.
Outcome:	The proposed framework outperforms data curation and training regularization on standard knowledge and reasoning benchmarks.

Copied to clipboard

Yixuan Su, Deng Cai, Qingyu Zhou, Zibo Lin, Simon Baker, Yunbo Cao, Shuming Shi, Nigel Collier, Yan Wang

Not All Preference Pairs Are Created Equal: A Recipe for Annotation-Efficient Iterative Preference Learning (2024.findings-emnlp)

Challenge:	Empirical studies on three benchmark datasets with three state-of-the-art matching models demonstrate that the proposed learning framework significantly improves the model performance across various evaluation metrics.
Approach:	They propose a hierarchical curriculum learning framework that trains matching models in an “easy-to-difficult” scheme.
Outcome:	The proposed framework significantly improves the model performance across evaluation metrics on three benchmark datasets with three state-of-the-art matching models.

Copied to clipboard

Sen Yang, Leyang Cui, Deng Cai, Xinting Huang, Shuming Shi, Wai Lam

Responsible & Trustworthy NLP

Challenge:	Large language models (LLMs) have shown remarkable capabilities to understand and generate human languages, supporting applications such as question answering, coding, and psychological counseling.
Approach:	They propose strategies to save annotation budgets while achieving competitive or even better performances for iterative preference learning.
Outcome:	The proposed methods save annotation budgets while achieving better performance.

Residual Learning of Neural Text Generation with n-gram Language Model (2022.findings-emnlp)

Copied to clipboard

Huayang Li, Deng Cai, Jin Xu, Taro Watanabe

RWKV-CLIP: A Robust Vision-Language Representation Learner (2024.emnlp-main)

Challenge:	Experimental results show that n-gram models can achieve satisfactory performance on a large proportion of testing cases.
Approach:	They propose to learn a neural LM that fits the residual between an n-gram LM and the real-data distribution.
Outcome:	The proposed model achieves additional performance gains over popular standalone models on three typical language tasks.

Copied to clipboard

Tiancheng Gu, Kaicheng Yang, Xiang An, Ziyong Feng, Dongnan Liu, Weidong Cai, Jiankang Deng

Towards Multi-dimensional Evaluation of LLM Summarization across Domains and Languages (2025.acl-long)

Challenge:	Using large image-text datasets, large-scale image-data sets have been used for visionlanguage pre-training.
Approach:	They propose a framework that leverages Large Language Models to combine and refine information from web-based image-text pairs, synthetic captions, and detection tags.
Outcome:	The proposed framework can combine and refine information from web-based image-text pairs, synthetic captions, and detection tags.

Copied to clipboard

Hyangsuk Min, Yuho Lee, Minjeong Ban, Jiaqi Deng, Nicole Hee-Yeon Kim, Taewon Yun, Hang Su, Jason Cai, Hwanjun Song

Summarization

Challenge:	Existing evaluation frameworks for text summarization lack domain-specific assessment criteria and are predominantly English-centric.
Approach:	They propose a multi-dimensional, multi-domain evaluation of summarization in English and Chinese that incorporates specialized assessment criteria for each domain and leverages a debate system to enhance annotation quality.
Outcome:	The proposed evaluation framework provides a multi-dimensional, multi-domain evaluation of summarization in English and Chinese.

OneRec-Think: In-Text Reasoning for Generative Recommendation (2026.acl-long)

Copied to clipboard

Zhanyu Liu, Shiyao Wang, Xingmei Wang, Rongzhou Zhang, Jiaxin Deng, Honghui Bao, Jinghao Zhang, Wuchao Li, PengFei Zheng, Xiangyu Wu, Yifei Hu, Qigen Hu, Xinchen Luo, Lejian Ren, Zhang Zixing, Qianqian Wang, Kuo Cai, Yunfan Wu, Hongtao Cheng, Zexuan Cheng, Lu Ren, Huanjie Wang, Yi Su, Ruiming Tang, Kun Gai, Guorui Zhou

Specialist or Generalist? Instruction Tuning for Specific NLP Tasks (2023.emnlp-main)

Challenge:	Existing generative models lack the capacity for explicit and controllable reasoning, a key advantage of LLMs.
Approach:	They propose a framework that integrates dialogue, reasoning, and personalized recommendation.
Outcome:	Experiments across public benchmarks show state-of-the-art performance.

Copied to clipboard

Chufan Shi, Yixuan Su, Cheng Yang, Yujiu Yang, Deng Cai

Linearizing Transformer with Key-Value Memory (2022.emnlp-main)

Challenge:	Recent studies have shown that instruction tuning can be a data-efficient method for transforming large language models into generalist models, but their performance lags behind specialist models trained exclusively for specific tasks.
Approach:	They propose to incorporate broadcoverage generalist instruction tuning into large language models to build a specialist model by incorporating task specificity and skill requirements.
Outcome:	The proposed method improves model performance when task coverage is broad and when training data is limited.

Copied to clipboard

Yizhe Zhang, Deng Cai

Large Language Models Meet Harry Potter: A Dataset for Aligning Dialogue Agents with Characters (2023.findings-emnlp)

Challenge:	Efficient transformer variants with linear time complexity have been developed to mitigate the quadratic computational overhead of the vanilla transformer.
Approach:	They propose a linear time complexity transformer variant that reduces the quadratic computational overhead of the vanilla transformer by using a recurrent-style incremental computation similar to kernel-based transformers.
Outcome:	The proposed method reduces the performance gap while achieving the same efficiency even with short generation.

Copied to clipboard

Nuo Chen, Yan Wang, Haiyun Jiang, Deng Cai, Yuhan Li, Ziyang Chen, Longyue Wang, Jia Li

Exploring Compositional Generalization of Multimodal LLMs for Medical Imaging (2025.acl-long)

Challenge:	Existing models that can create open-domain dialogue agents lack character representation and annotations.
Approach:	They propose a dataset to study character alignment and character representation . it includes all dialogue sessions from the Harry Potter series and includes annotations .
Outcome:	The proposed dataset can be used as a universal benchmark for character-driven LLMs.

Copied to clipboard

Zhenyang Cai, Junying Chen, Rongsheng Wang, Weihong Wang, Yonglin Deng, Dingjie Song, Yize Chen, Zixu Zhang, Benyou Wang