Papers by Hao Cheng

Adaptive Spatial and Temporal Redundancy Optimization for Efficient Reasoning in Large Language Models (2026.acl-long)

Challenge:	Existing domain-specific knowledge of domain-related tasks is lacking in pre-trained language models.
Approach:	They propose a domain-adaptation method which can dynamically select domain-specific tokens and guide the discriminator to emphasize them, without introducing new training parameters.
Outcome:	The proposed method can capture domain-specific knowledge of domain-related tasks without introducing new training parameters.

Copied to clipboard

Tianle Chen, Pengyu Cheng, Qiyuan Zhu, Jiacheng Wang, Bei Liu, Hao Gu, Ruijie Shen, Xiaofeng Hou, Sirui Han, Jiacheng Liu

Language Models as Inductive Reasoners (2024.eacl-long)

Challenge:	Existing research to improve CoT efficiency falls into three categories, each with distinct limitations.
Approach:	They propose a training-free framework that addresses both dimensions of CoT reasoning by applying a progressive precision reduction strategy coupled with an entropy-based confidence mechanism for adaptive termination.
Outcome:	Empirical results show that the proposed framework achieves 11.3 efficiency gain without compromising accuracy.

Copied to clipboard

Zonglin Yang, Li Dong, Xinya Du, Hao Cheng, Erik Cambria, Xiaodong Liu, Jianfeng Gao, Furu Wei

Pre-training Multi-task Contrastive Learning Models for Scientific Literature Understanding (2023.findings-emnlp)

Challenge:	Inductive reasoning is a core component of human intelligence.
Approach:	They propose a task to induce natural language rules from natural language facts using natural language as representation for knowledge instead of formal language.
Outcome:	The proposed task surpasses baselines in both automatic and human evaluations.

Copied to clipboard

Yu Zhang, Hao Cheng, Zhihong Shen, Xiaodong Liu, Ye-Yi Wang, Jianfeng Gao

InSCIt: Information-Seeking Conversations with Mixed-Initiative Interactions (2023.tacl-1)

Challenge:	Pre-trained language models (LMs) have shown effectiveness in literature understanding tasks, especially when tuned via contrastive learning.
Approach:	They propose a multi-task contrastive learning framework that enables common knowledge sharing across different scientific literature understanding tasks while preventing task-specific skills from interfering with each other.
Outcome:	The proposed framework outperforms state-of-the-art pre-trained language models on a comprehensive dataset.

Copied to clipboard

Zeqiu Wu, Ryu Parish, Hao Cheng, Sewon Min, Prithviraj Ammanabrolu, Mari Ostendorf, Hannaneh Hajishirzi

SafetyQuizzer: Timely and Dynamic Evaluation on the Safety of LLMs (2025.naacl-long)

Challenge:	In information-seeking conversations, a user may ask questions that are under-specified or unanswerable.
Approach:	They present a dataset for information-seeking conversations with mixed-initiative interactions . they use Wikipedia to search for answers and provide relevant information .
Outcome:	The proposed system significantly underperforms humans in two of the most recent studies.

Copied to clipboard

Zhichao Shi, Shaoling Jing, Yi Cheng, Hao Zhang, Yuanzhuo Wang, Jie Zhang, Huawei Shen, Xueqi Cheng

A Self-supervised Joint Training Framework for Document Reranking (2022.findings-naacl)

Challenge:	Large Language Models (LLMs) have been used to evaluate the safety of their users . however, evaluation questions in current benchmarks are too straightforward and difficult to update with practical relevance due to their lack of correlation with real-world events.
Approach:	They propose a question-generation framework to evaluate the safety of LLMs in the Chinese context.
Outcome:	The proposed framework reduces decline rate while maintaining similar attack success rate.

Copied to clipboard

Xiaozhi Zhu, Tianyong Hao, Sijie Cheng, Fu Lee Wang, Hai Liu

ChartM3: A Multi-Stage Code-Driven Pipeline for Constructing Multi-Dimensional and Multi-Step Visual Reasoning Data in Chart Comprehension (2025.findings-emnlp)

Challenge:	Pretrained language models have been successfully applied to a wide range of tasks . however, the pretraining tasks were based on the context of documents .
Approach:	They propose a self-supervised joint training framework with a method called Masked Query Prediction to establish semantic relations between given queries and positive documents.
Outcome:	The proposed framework outperforms existing models on document reranking tasks without further pre-training . it uses a self-supervised method to establish semantic relations between given queries and positive documents.

Copied to clipboard

Duo Xu, Hao Cheng, Xin Lin, Zhen Xie, Hao Henry Wang

DocLens: Multi-aspect Fine-grained Medical Text Evaluation (2024.acl-long)

Challenge:	Currently, research on complex chart understanding tasks is limited . a pipeline for visual reasoning datasets addresses these limitations .
Approach:	They propose a code-driven pipeline for generating visual reasoning datasets . pipeline integrates retrieval-augmented generation to retrieve professional chart templates .
Outcome:	The proposed pipeline enhances chart diversity and data quality through model-based evaluation.

Copied to clipboard

Yiqing Xie, Sheng Zhang, Hao Cheng, Pengfei Liu, Zelalem Gero, Cliff Wong, Tristan Naumann, Hoifung Poon, Carolyn Rose

Text Generation

Challenge:	Medical text generation systems are widely used to assist with administrative work and highlight salient information to support decision-making.
Approach:	They propose a set of metrics to evaluate completeness, conciseness, and attribution of medical text at a fine-grained level.
Outcome:	The proposed framework exhibits substantially higher agreement with medical experts than existing metrics.

Shifting Attention to Relevance: Towards the Predictive Uncertainty Quantification of Free-Form Large Language Models (2024.acl-long)

Copied to clipboard

Jinhao Duan, Hao Cheng, Shiqi Wang, Alex Zavalny, Chenan Wang, Renjing Xu, Bhavya Kailkhura, Kaidi Xu

Unsupervised Learning of Hierarchical Conversation Structure (2022.findings-emnlp)

Challenge:	Large Language Models (LLMs) show promising results in language generation but often “hallucinate”, making their outputs less reliable.
Approach:	They propose to shift attention to more relevant components at token- and sentence-levels for better UQ.
Outcome:	The proposed approach improves the performance of a range of popular “off-the-shelf” LLMs with model sizes extending up to 33B parameters.

Copied to clipboard

Bo-Ru Lu, Yushi Hu, Hao Cheng, Noah A. Smith, Mari Ostendorf

The Microsoft Toolkit of Multi-Task Deep Neural Networks for Natural Language Understanding (2020.acl-demos)

Challenge:	Goal-oriented conversations often have sub-dialogue structure, but it can be domain-dependent . Increasingly, language understanding applications involve conversational speech and text .
Approach:	They propose an unsupervised approach to learning hierarchical conversation structure . they use turn and sub-dialogue segment labels to decode the structure based on dialogue acts and subtasks .
Outcome:	The proposed approach improves neural models for three conversation-level understanding tasks.

Copied to clipboard

Xiaodong Liu, Yu Wang, Jianshu Ji, Hao Cheng, Xueyun Zhu, Emmanuel Awa, Pengcheng He, Weizhu Chen, Hoifung Poon, Guihong Cao, Jianfeng Gao

ReEval: Automatic Hallucination Evaluation for Retrieval-Augmented Large Language Models via Transferable Adversarial Attacks (2024.findings-naacl)

Challenge:	MT-DNN is an open-source natural language understanding toolkit . it allows researchers and developers to train customized deep learning models .
Approach:	They present MT-DNN, an open-source natural language understanding toolkit . it is designed to facilitate rapid customization for a broad spectrum of NLU tasks . MT supports multi-task knowledge distillation, which can substantially compress a deep neural model without significant performance drop.
Outcome:	The proposed model can significantly compress a large model without significant performance drop.

Copied to clipboard

Xiaodong Yu, Hao Cheng, Xiaodong Liu, Dan Roth, Jianfeng Gao

Data Mixing Agent: Learning to Re-weight Domains for Continual Pre-training (2026.acl-long)

Challenge:	Existing static benchmarks do not guarantee that models can use the provided evidence for answering, which is essential to avoid hallucination when the required knowledge is new or private.
Approach:	They propose to automatically perturb existing static one for dynamic evaluation by using a chatGPT framework and a set of open-domain QA datasets.
Outcome:	The proposed framework generates new test cases on two open-domain QA datasets and is human-readable and useful to trigger hallucination in LLMs.

Copied to clipboard

Kailai Yang, Xiao Liu, Lei Ji, Hao Li, Xiao Liang, Zhiwei Liu, Yeyun Gong, Peng Cheng, Mao Yang

Knowledge-enhanced Prompt Tuning for Dialogue-based Relation Extraction with Trigger and Label Semantic (2024.lrec-main)

Challenge:	Existing methods for reweighting data mixtures rely on manual designation with certain heuristics based on intuition or empirical results.
Approach:	They propose a model-based framework that learns to re-weight domains by reinforcement learning on large quantities of data mixing trajectories with corresponding feedback from an evaluation environment.
Outcome:	The proposed framework outperforms baselines in achieving balanced performance across source and target fields and domain spaces without retraining.

Copied to clipboard

Hao An, Zhihong Zhu, Xuxin Cheng, Zhiqi Huang, Yuexian Zou

Train Once, and Decode As You Like (2020.coling-main)

Challenge:	Existing methods to determine semantic relation between two arguments in dialogues are limited due to the low information density of text.
Approach:	They propose a Knowledge-Enhanced Prompt-Tuning method to enhance DRE model by exploiting trigger and label semantics.
Outcome:	The proposed method achieves state-of-the-art in F1 and F1c scores on a DialogRE dataset.

Copied to clipboard

Chao Tian, Yifei Wang, Hao Cheng, Yijiang Lian, Zhihua Zhang

Text Generation

Challenge:	Existing approaches to machine translation support autoregressive, semi-autoregressive and refinement-based non-auto-regressives.
Approach:	They propose a unified approach for supporting different generation manners of machine translation including autoregressive, semi-autoregressive and refinement-based non-auto-regressives.
Outcome:	The proposed approach achieves better or competitive translation performance compared with strong baseline models in all the settings.

Novel Relation Detection: Discovering Unknown Relation Types via Multi-Strategy Self-Supervised Learning (2023.findings-emnlp)

Copied to clipboard

Qingbin Liu, Yin Kung, Yanchao Hao, Dianbo Sui, Siyuan Cheng, Xi Chen, Ningyu Zhang, Jiaoyan Chen

Low-Resource NLP

Challenge:	Existing approaches to relation extraction can only recognize predefined relation types . new or out-of-scope relation types may continually emerge after the model is deployed .
Approach:	They propose a novel relation detection task that uses self-supervised learning to handle shallow semantic similarity problem.
Outcome:	The proposed method outperforms state-of-the-art methods on two datasets.

InstructSafety: A Unified Framework for Building Multidimensional and Explainable Safety Detector through Instruction Tuning (2023.findings-emnlp)

Copied to clipboard

Zhexin Zhang, Jiale Cheng, Hao Sun, Jiawen Deng, Minlie Huang

Vision-Language Models Can Self-Improve Reasoning via Reflection (2025.naacl-long)

Challenge:	Existing safety detection systems have limitations in terms of their versatility and interpretability.
Approach:	They introduce a safety detection framework that unifies 7 common sub-tasks into a uniform formulation and process 39 human-annotated datasets for instruction tuning.
Outcome:	The proposed framework unifies 7 common sub-tasks into a uniform formulation and then runs on 39 human-annotated datasets to fine-tune it.

Copied to clipboard

Kanzhi Cheng, Li YanTao, Fangzhi Xu, Jianbing Zhang, Hao Zhou, Yang Liu

Iterative Structured Pruning for Large Language Models with Multi-Domain Calibration (2026.eacl-industry)

Challenge:	Chain-of-thought (CoT) has been shown to improve the reasoning capability of large language models (LLMs).
Approach:	They propose a framework which iteratively enhances the model’s Vision-language Reasoning by Reflecting on CoT Rationales.
Outcome:	The proposed framework improves multimodal reasoning on vision-language tasks by 23% to 60% over baselines.

Copied to clipboard

Guangxin Wu, Hao Zhang, Zhang Zhibin, Jiafeng Guo, Xueqi Cheng

Constructing Highly Inductive Contexts for Dialogue Safety through Controllable Reverse Generation (2022.findings-emnlp)

Challenge:	Existing models with unstructured pruning often yield irregular sparsity patterns that necessitate specialized hardware or software support.
Approach:	They propose a structured pruning framework that eliminates entire architectural components and maintains compatibility with standard hardware accelerators.
Outcome:	The proposed model pruning framework achieves significant compression with minimal performance degradation on multiple models across diverse downstream tasks.

Copied to clipboard

Zhexin Zhang, Jiale Cheng, Hao Sun, Jiawen Deng, Fei Mi, Yasheng Wang, Lifeng Shang, Minlie Huang

Iterative Self-Tuning LLMs for Enhanced Jailbreaking Capabilities (2025.naacl-long)

Challenge:	Existing methods to detect toxic generation of pretrained language models rely on templates, data extraction, crowdsourcing workers or automatic generation.
Approach:	They propose a method to construct adversarial contexts conditioned on a given response . they augment existing dataset BAD+ and construct a new dataset B AD+ .
Outcome:	The proposed method can detect toxic or biased content in large pretrained language models.

Copied to clipboard

Chung-En Sun, Xiaodong Liu, Weiwei Yang, Tsui-Wei Weng, Hao Cheng, Aidan San, Michel Galley, Jianfeng Gao

Robustness in NLP

Challenge:	Recent research shows that Large Language Models (LLMs) are vulnerable to automated jailbreak attacks.
Approach:	They propose a framework that crafts adversarial LLMs with enhanced jailbreak ability.
Outcome:	ADV-LLM significantly reduces the computational cost of generating adversarial suffixes while achieving nearly 100% ASR on various open-source LLMs.

DualRAG: A Dual-Process Approach to Integrate Reasoning and Retrieval for Multi-Hop Question Answering (2025.acl-long)

Copied to clipboard

Rong Cheng, Jinyi Liu, Yan Zheng, Fei Ni, Jiazhen Du, Hangyu Mao, Fuzheng Zhang, Bo Wang, Jianye Hao

Con-NAT: Contrastive Non-autoregressive Neural Machine Translation (2022.findings-emnlp)

Challenge:	Existing approaches to multi-hop question answering struggle to identify and organize dynamic knowledge . et al., 2023; Liu e.t. al. 2023) suggest a dual-process framework for multi-step reasoning .
Approach:	They propose a synergistic dual-process framework that integrates reasoning and retrieval.
Outcome:	The proposed framework improves answer accuracy and coherence even in smaller-scale models.

Copied to clipboard

Hao Cheng, Zhihua Zhang

Task-Aware Specialization for Efficient and Robust Dense Retrieval for Open-Domain Question Answering (2023.acl-short)

Challenge:	Neural machine translation models are autoregressive, which means they predict tokens one by one based on source tokens and previously predicted tokens.
Approach:	They propose a conditional masked language model which incorporates contrastive learning into the conditional language model.
Outcome:	The proposed model improves on WMT’16 Ro-En translation directions with different data sizes.

Copied to clipboard

Hao Cheng, Hao Fang, Xiaodong Liu, Jianfeng Gao

Tackling Distractor Documents in Multi-Hop QA with Reinforcement and Curriculum Learning (2026.findings-eacl)

Challenge:	Existing dense retrieval models are parameter-inefficient and underperform sparse counterparts.
Approach:	They propose a task-aware specialization for dEnse Retrieval architecture that enables parameter sharing by interleaving shared and specialized blocks in a single encoder.
Outcome:	The proposed architecture surpasses BM25 on questions and passages using 60% of the parameters as bi-encoder dense retrievers.

Copied to clipboard

Jerry Huang, Siddarth Madala, Risham Sidhu, Cheng Niu, Hao Peng, Julia Hockenmaier, Tong Zhang

Information Retrieval

Challenge:	Existing work on retrieval-augmented generation systems has shown that retrievers exhibit imperfect recall and precision, limiting downstream performance.
Approach:	They propose a retrieval-augmented generation model that generates answers from larger sets of retrieved contexts.
Outcome:	The proposed model generates answers and cites relevant information from larger sets of retrieved contexts.

MARCH: Multi-Agent Reinforced Check for Hallucination (2026.acl-long)

Copied to clipboard

Zhuo Li, Yupeng Zhang, Pengyu Cheng, Jiajun Song, Mengyu Zhou, Hao Li, Shujie Hu, Yu Qin, null Erchao.zec, Xiaoxi Jiang, null Guanjunjiang

Reasoning

Challenge:	Existing methods to detect hallucinations suffer from inherent confirmation bias, where the verifier inadvertently reproduces the errors of the original generation.
Approach:	They propose a framework that enforces rigorous factual alignment by leveraging deliberate information asymmetry by combining a pipeline of three specialized agents: a Solver, a Proposer, and a Checker.
Outcome:	Extensive experiments across hallucination benchmarks demonstrate that MARCH substantially reduces hallucinism rates.

MPL: Multiple Programming Languages with Large Language Models for Information Extraction (2025.findings-acl)

Copied to clipboard

Bo Li, Gexiang Fang, Wei Ye, Zhenghua Xu, Jinglei Zhang, Hao Cheng, Shikun Zhang

On the Safety of Conversational Models: Taxonomy, Dataset, and Benchmark (2022.findings-acl)

Challenge:	Existing research focuses on Python for code-style simulation, overlooking the potential of other widely-used PLs during the supervised fine-tuning phase.
Approach:	They propose a framework that incorporates programming languages into IE tasks . they introduce function-prompt with virtual running to simulate code-style inputs .
Outcome:	The proposed framework exploits the potential of different programming languages during the supervised fine-tuning phase.

Copied to clipboard

Hao Sun, Guangxuan Xu, Jiawen Deng, Jiale Cheng, Chujie Zheng, Hao Zhou, Nanyun Peng, Xiaoyan Zhu, Minlie Huang

OrchestraLLM: Efficient Orchestration of Language Models for Dialogue State Tracking (2024.naacl-long)

Challenge:	Dialogue safety problems severely limit the real-world deployment of generative conversational models.
Approach:	They propose a taxonomy for dialogue safety specifically designed to capture unsafe behaviors in human-bot dialogue settings.
Outcome:	The proposed taxonomy captures unsafe behaviors in human-bot dialogue settings with rich context-sensitive unsafe examples.

Copied to clipboard

Chia-Hsuan Lee, Hao Cheng, Mari Ostendorf

PAL: Persona-Augmented Emotional Support Conversation Generation (2023.findings-acl)

Challenge:	Large language models (LLMs) are computationally expensive and often require computational resources.
Approach:	They propose a routing framework that seamlessly integrates a SLM and an LLM, or-lm, or a LLM into a single framework.
Outcome:	The proposed routing framework reduces the computational costs by over 50% in dialogue state tracking tasks.

Copied to clipboard

Jiale Cheng, Sahand Sabour, Hao Sun, Zhuang Chen, Minlie Huang

StableToolBench-MirrorAPI: Modeling Tool Environments as Mirrors of 7,000+ Real-World APIs (2025.findings-acl)

Challenge:	Recent work has demonstrated the effectiveness of dialogue models in providing emotional support due to the lack of human resources for mental health support.
Approach:	They propose a framework for dynamically inferring and modeling seekers’ persona from the conversation history and a model that leverages persona information to provide personalized emotional support.
Outcome:	The proposed model outperforms baseline models on the studied benchmark.

Copied to clipboard

Zhicheng Guo, Sijie Cheng, Yuchen Niu, Hao Wang, Sicheng Zhou, Wenbing Huang, Yang Liu

Knowledge-Rich Self-Supervision for Biomedical Entity Linking (2022.findings-emnlp)

Challenge:	Existing tool environments face challenges in balancing stability, scale, and realism, especially for benchmarking purposes.
Approach:	They propose a framework that trains specialized LLMs to accurately simulate real API responses by supervised fine-tuning and chain-of-thought reasoning.
Outcome:	The proposed framework achieves superior accuracy and stability compared to state-of-the-art methods on the newly constructed MirrorAPI-Bench and its integration into StableToolBench.

Copied to clipboard

Sheng Zhang, Hao Cheng, Shikhar Vashishth, Cliff Wong, Jinfeng Xiao, Xiaodong Liu, Tristan Naumann, Jianfeng Gao, Hoifung Poon

LEDGER: Scaling Agentic Document Editing with Dependency-aware Graph Retrieval (2026.findings-acl)

Challenge:	Entity linking is challenging in high-value domains with myriad entities . standard classification approaches suffer from the annotation bottleneck .
Approach:	They propose a self-supervised approach to learn domain knowledge for biomedical entity linking . it generates self-reported mention examples on unlabeled text and trains contextual encoder .
Outcome:	The proposed method outperforms existing methods by 20 points in accuracy on biomedical datasets.

Copied to clipboard

Hang Wang, Utkarsh Garg, Reza Davari, Huitian Jiao, Hao Cheng, Baolin Peng, Si-Qing Chen, Tao Ge

Green & Sustainable NLP

Challenge:	Document editing requires full-context awareness of dependencies, but processing entire documents for each edit incurs prohibitive token costs and latency.
Approach:	a framework that constructs lightweight dependency graphs captures semantic relationships and structural hierarchies across document elements is proposed for agentic document editing . a scaLing agentic agentic framework is based on a dependency graph framework that captures dependencies and refactors function dependencies.
Outcome:	a new framework achieves 76 consistency versus 56 baseline while reducing token usage by 85 . the framework is based on a framework that captures semantic relationships and structural hierarchies across document elements . it can be used to improve document consistency, but it also reduces token costs and latency .

Chain-of-Skills: A Configurable Model for Open-Domain Question Answering (2023.acl-long)

Copied to clipboard

Kaixin Ma, Hao Cheng, Yu Zhang, Xiaodong Liu, Eric Nyberg, Jianfeng Gao

Information Retrieval

Challenge:	Using customized retrieval models, model transferability and scalability are limited.
Approach:	They propose a modular retrieval model where individual modules correspond to key skills that can be reused across datasets.
Outcome:	The proposed model outperforms self-supervised retrievers in zero-shot evaluations and achieves state-of-the-art fine-tuned retrieval performance on NQ, HotpotQA and OTT-QA.

AndroidLab: Training and Systematic Benchmarking of Android Autonomous Agents (2025.acl-long)

Copied to clipboard

Yifan Xu, Xiao Liu, Xueqiao Sun, Siyi Cheng, Hao Yu, Hanyu Lai, Shudan Zhang, Dan Zhang, Jie Tang, Yuxiao Dong

UnitedQA: A Hybrid Approach for Open Domain Question Answering (2021.acl-long)

Challenge:	Existing studies on Android agents lack systematic research on open-source and closed-source models.
Approach:	They propose a framework for Android agents that includes an operation environment and a reproducible benchmark.
Outcome:	The proposed framework lifts the success rate of open-source LLMs and LMMs from 4.59% to 21.50% for LLM and 1.93% to 13.28% for LMM.

Copied to clipboard

Hao Cheng, Yelong Shen, Xiaodong Liu, Pengcheng He, Weizhu Chen, Jianfeng Gao

StableToolBench: Towards Stable Large-Scale Benchmarking on Tool Learning of Large Language Models (2024.findings-acl)

Challenge:	Recent work on open-domain question answering focuses on either extractive or generative readers exclusively.
Approach:	They propose a hybrid approach to extractive and generative readers that leverages both models.
Outcome:	The proposed approach outperforms state-of-the-art models on NaturalQuestions and TriviaQA respectively.

Copied to clipboard

Zhicheng Guo, Sijie Cheng, Hao Wang, Shihao Liang, Yujia Qin, Peng Li, Zhiyuan Liu, Maosong Sun, Yang Liu

Open-domain Question Answering via Chain of Reasoning over Heterogeneous Knowledge (2022.findings-emnlp)

Challenge:	Large Language Models (LLMs) have witnessed remarkable advancements in recent years, prompting the exploration of tool learning.
Approach:	They propose a virtual API server and stable evaluation system to assess the stability of large-scale real-time APIs.
Outcome:	The proposed benchmarks demonstrate the stability of the proposed system and its caching system.

Copied to clipboard

Kaixin Ma, Hao Cheng, Xiaodong Liu, Eric Nyberg, Jianfeng Gao

Reasoning

Challenge:	Existing open-domain question answering methods rely on the retriever to gather all evidence in isolation, but our approach uses an intermediary module to perform a chain of reasoning over the retrieved set.
Approach:	They propose a new open-domain question answering framework that integrates an intermediary module into the current retriever-reader pipeline and integrates it into the model.
Outcome:	The proposed framework outperforms the state-of-the-art on two OTT-QA datasets with an exact match score of 47.3 (45% relative gain).

Zero-Shot Spoken Language Understanding via Large Language Models: A Preliminary Study (2024.lrec-main)

Copied to clipboard

Zhihong Zhu, Xuxin Cheng, Hao An, Zhichang Wang, Dongsheng Chen, Zhiqi Huang

TLUE: A Tibetan Language Understanding Evaluation Benchmark (2025.emnlp-main)

Challenge:	Recent advances in large language models (LLMs) have shown promising results in zero-shot settings, which motivates us to explore prompt-based methods.
Approach:	They propose a two-stage framework which transforms the SLU task into a question-answering problem by directly prompting LLMs.
Outcome:	The proposed framework can be built by directly prompting LLMs to understand user needs without training data.

Copied to clipboard

Fan Gao, Cheng Huang, Yutong Liu, Nyima Tashi, Xiangxiang Wang, Thupten Tsering, Ban Ma-bao, Renzeng Duojie, Gadeng Luosang, Rinchen Dongrub, Dorje Tashi, Xiao Feng Cd, Yongbin Yu, Hao Wang

OpenCoder: The Open Cookbook for Top-Tier Code Large Language Models (2025.acl-long)

Challenge:	Low-resource languages, like Tibetan, remain underrepresented in large language models' evaluations.
Approach:	They propose a Tibetan Language Understanding Evaluation Benchmark to assess LLMs' proficiency in Tibetan . they use a multi-task understanding benchmark and a safety benchmark to evaluate models .
Outcome:	The proposed benchmark shows that most large language models perform below the random baseline, especially in Tibetan language processing.

Copied to clipboard

Siming Huang, Tianhao Cheng, Jason Klein Liu, Weidi Xu, Jiaran Hao, Liuyihan Song, Yang Xu, Jian Yang, Jiaheng Liu, Chenchen Zhang, Linzheng Chai, Ruifeng Yuan, Xianzhen Luo, Qiufeng Wang, YuanTao Fan, Qingfu Zhu, Zhaoxiang Zhang, Yang Gao, Jie Fu, Qian Liu, Houyi Li, Ge Zhang, Yuan Qi, Xu Yinghui, Wei Chu, Zili Wang

Dialogue State Tracking with a Language Model using Schema-Driven Prompting (2021.emnlp-main)

Challenge:	Code LLMs lack reproducible data pipelines and training protocols for reproducible advancements in code intelligence.
Approach:	They propose a top-tier code LLM that releases model weights and inference code . reproducible data pipelines, rigorous experimental ablation results and training protocols are included .
Outcome:	The proposed model achieves comparable performance to leading models and serves as an "open cookbook" reproducible training data, rigorous experimental ablation results, and detailed training protocols are also included in the model.

Copied to clipboard

Chia-Hsuan Lee, Hao Cheng, Mari Ostendorf

AceGPT, Localizing Large Language Models in Arabic (2024.naacl-long)

Challenge:	Task-oriented conversational systems often use dialogue state tracking to represent the user’s intentions, which involves filling in values of pre-defined slots.
Approach:	They propose a schema-driven prompting approach that provides task-aware history encoding that is used for both categorical and non-categorical slots.
Outcome:	The proposed system achieves state-of-the-art performance on MultiWOZ 2.2 and competitive performance on two other benchmarks: MultiWOz 2.1 and M2M.

Copied to clipboard

Huang Huang, Fei Yu, Jianqing Zhu, Xuening Sun, Hao Cheng, Song Dingjie, Zhihong Chen, Mosen Alharthi, Bang An, Juncai He, Ziche Liu, Junying Chen, Jianquan Li, Benyou Wang, Lian Zhang, Ruoyu Sun, Xiang Wan, Haizhou Li, Jinchao Xu

Posterior Differential Regularization with f-divergence for Improving Model Robustness (2021.naacl-main)

Challenge:	Significant concerns emerge when addressing cultural sensitivity and local values.
Approach:	They propose a localized Large Language Model (LLM) specifically for Arabic, a language imbued with unique cultural characteristics inadequately addressed by current mainstream models.
Outcome:	The proposed model sets the state-of-the-art standard for open Arabic LLMs across various benchmarks.

Copied to clipboard

Hao Cheng, Xiaodong Liu, Lis Pereira, Yaoliang Yu, Jianfeng Gao

Robustness in NLP

Challenge:	Recent studies show that pre-trained models suffer catastrophic degradation in out-of-domain generalization to datasets with domain shift or adversarial scenarios.
Approach:	They propose to regularize the posterior difference between clean and noisy inputs by using a Jacobian regularization framework and a virtual adversarial training framework.
Outcome:	The proposed framework can improve model robustness in fully supervised and semi-supervised settings.

Explainable Recommendation with Personalized Review Retrieval and Aspect Learning (2023.acl-long)

Copied to clipboard

Hao Cheng, Shuo Wang, Wensheng Lu, Wei Zhang, Mingyang Zhou, Kezhong Lu, Hao Liao

Explainability & Interpretability in NLP

Challenge:	Recent years have witnessed a growing interest in the development of explainable recommendation models.
Approach:	They propose a model that combines prediction and generation tasks to produce more persuasive explanations by obtaining additional information from the training sets.
Outcome:	The proposed model outperforms state-of-the-art models on three datasets and shows that it is more persuasive than previous models.

Modelling Long-distance Node Relations for KBQA with Global Dynamic Graph (2020.coling-main)

Copied to clipboard

Xu Wang, Shuai Zhao, Jiale Han, Bo Cheng, Hao Yang, Jianchang Ao, Zhenzi Li

Structured Data in NLP

Challenge:	Existing studies rely on deep graph neural networks (GNNs) to capture rich structural information, but they lack the structural information needed for QA.
Approach:	They propose a framework which captures structural information from KBs and models long-distance node relations from two perspectives.
Outcome:	The proposed framework models long-distance node relations from two perspectives . it is based on two widely used multi-hop KBQA datasets .

PRISM-MCTS: Learning from Reasoning Trajectories with Metacognitive Reflection (2026.findings-acl)

Copied to clipboard

Siyuan Cheng, Bozhong Tian, Yanchao Hao, Zheng Wei

Sounding Board: A User-Centric and Content-Driven Social Chatbot (N18-5)

Challenge:	Existing reasoning models are limited by inefficiency and computational redundancy . PRISM-MCTS integrates a process reward model with a dynamic shared memory .
Approach:	They propose a reasoning framework that integrates a process reward model with a dynamic shared memory.
Outcome:	PRISM-MCTS integrates a process reward model with a dynamic shared memory . it halves trajectory requirements on GPQA while surpassing MCTS-RAG and Search-o1 .

Copied to clipboard

Hao Fang, Hao Cheng, Maarten Sap, Elizabeth Clark, Ari Holtzman, Yejin Choi, Noah A. Smith, Mari Ostendorf

Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models (2026.acl-long)

Challenge:	Sounding Board is a social chatbot that can hold a coherent conversation with humans . the system is user-centric in that users can control the topic of conversation, while the system adapts to the user's needs.
Approach:	They present Sounding Board, a social chatbot that won the 2017 Amazon Alexa Prize.
Outcome:	The system is user-centric in that users can control the topic of conversation, while the system adapts to the user's needs.

Copied to clipboard

Xin Cheng, Wangding Zeng, Damai Dai, Qinyu Chen, Bingxuan Wang, Zhenda Xie, Kezhao Huang, Xingkai Yu, Zhewen Hao, Han Zhang, Yu-Kun Li, Huishuai Zhang, Dongyan Zhao, Wenfeng Liang

UPRISE: Universal Prompt Retrieval for Improving Zero-Shot Evaluation (2023.emnlp-main)

Challenge:	Mixture-of-Experts (MoE) scales capacity via conditional computation, but lacks knowledge lookup primitive.
Approach:	They propose a conditional memory instantiated via Deep Sparse Embedding (DSE) they propose 'u-shaped scaling law' that identifies optimal balance between MoE experts and DSE memory .
Outcome:	The proposed model outperforms an iso-parameter and isoFLOPs MoE baseline across knowledge and reasoning benchmarks and is infrastructure-efficient.

Copied to clipboard

Daixuan Cheng, Shaohan Huang, Junyu Bi, Yuefeng Zhan, Jianfeng Liu, Yujing Wang, Hao Sun, Furu Wei, Weiwei Deng, Qi Zhang

Soul-Mix: Enhancing Multimodal Machine Translation with Manifold Mixup (2024.acl-long)

Challenge:	Large Language Models (LLMs) have impressive capabilities but need for task-specific prompt engineering can hinder their generalization.
Approach:	They propose a lightweight and versatile retriever that automatically retrieves prompts for a given zero-shot task input.
Outcome:	The proposed model is universally applicable across tasks and models . it mitigates hallucination problem in chatGPT, and it improves even the strongest LLMs.

Copied to clipboard

Xuxin Cheng, Ziyu Yao, Yifei Xin, Hao An, Hongxiang Li, Yaowei Li, Yuexian Zou

Open Domain Question Answering with A Unified Knowledge Interface (2022.acl-long)

Challenge:	Multimodal machine translation (MMT) aims to improve the performance of machine translation with the help of visual information.
Approach:	They propose a multimodal machine translation mixup method that integrates visual information into conventional text-only neural machine translation systems.
Outcome:	The proposed method outperforms existing models on a multi-directional dataset with fewer parameters and achieves new state-of-the-art performance.

Copied to clipboard

Kaixin Ma, Hao Cheng, Xiaodong Liu, Eric Nyberg, Jianfeng Gao

Structured Data in NLP

Challenge:	a retriever-reader framework is popular for open domain question answering . however, accessing heterogeneous knowledge sources through a unified interface remains unknown .
Approach:	They propose a retriever-reader framework that uses explicit knowledge to access heterogeneous knowledge sources through a unified interface.
Outcome:	The proposed framework can benefit from the expanded knowledge index, the authors show . their approach sets the single-model state-of-the-art on Natural Questions .

Shallow Focus, Deep Fixes: Enhancing Shallow Layers Vision Attention Sinks to Alleviate Hallucination in LVLMs (2025.emnlp-main)

Copied to clipboard

Xiaofeng Zhang, Yihao Quan, Chen Shen, Chaochen Gu, Xiaosong Yuan, Shaotian Yan, Jiawei Cao, Hao Cheng, Kaijie Wu, Jieping Ye

A Dynamic Speaker Model for Conversational Interactions (N19-1)

Challenge:	Multimodal large language models (MLLMs) demonstrate excellent abilities for understanding visual information, but the hallucination remains a challenging problem.
Approach:	They propose a training-free approach to enhance vision attention sinks to facilitate convergence of the image token attention sink within shallow layers.
Outcome:	The proposed approach improves the convergence of the image token attention sink within shallow layers and strengthens the layer’s focus on the image itself.

Copied to clipboard

Hao Cheng, Hao Fang, Mari Ostendorf

Targeted Adversarial Training for Natural Language Understanding (2021.naacl-main)

Challenge:	a neural model for characterizing individual differences in speakers is shown to be useful in human-computer interaction and dialog act prediction.
Approach:	They propose a neural model for learning a dynamically updated speaker embedding in a conversational context.
Outcome:	The proposed model is used for content ranking and dialog act prediction in human-human conversations.

Copied to clipboard

Lis Pereira, Xiaodong Liu, Hao Cheng, Hoifung Poon, Jianfeng Gao, Ichiro Kobayashi

Robustness in NLP

Challenge:	Existing adversarial training approaches focus on making adversarials less expensive or regularizing rather than replacing the standard training objective.
Approach:	They propose an algorithm to introspect current mistakes and prioritize adversarial training steps to where the model errs the most.
Outcome:	The proposed algorithm improves adversarial training for natural language understanding by introspecting mistakes and prioritizing training steps to where the model errs the most.

Diversifying the Expert Knowledge for Task-Agnostic Pruning in Sparse Mixture-of-Experts (2025.findings-acl)

Copied to clipboard

Zeliang Zhang, Xiaodong Liu, Hao Cheng, Chenliang Xu, Jianfeng Gao

Challenge:	Large Language Models (LLMs) have outstanding performance by learning a large number of model parameters on large amounts of data.
Approach:	They propose a method of grouping and pruning similar experts to improve the model’s parameter efficiency by a range of natural language tasks.
Outcome:	The proposed method outperforms other model pruning methods on a range of natural language tasks.

Probabilistic Assumptions Matter: Improved Models for Distantly-Supervised Document-Level Question Answering (2020.acl-main)

Copied to clipboard

Hao Cheng, Ming-Wei Chang, Kenton Lee, Kristina Toutanova