Papers by Haotian Zhang
Improve Vision Language Model Chain-of-thought Reasoning (2025.acl-long)
Copied to clipboard
Ruohong Zhang, Bowen Zhang, Yanghao Li, Haotian Zhang, Zhiqing Sun, Zhe Gan, Yinfei Yang, Ruoming Pang, Yiming Yang
| Challenge: | Current training recipes often rely on datasets dominated by short annotations with limited rationales, hindering the models' ability to generalize to tasks requiring comprehensive reasoning. |
| Approach: | They propose a two-stage post-training strategy that augments short answers with CoT reasoning generated by GPT-4o, enhancing the VLM's CoT capabilities through fine-tuning. |
| Outcome: | The proposed strategy enhances the model's CoT capabilities through fine-tuning and reinforcement learning. |
E-NER: Evidential Deep Learning for Trustworthy Named Entity Recognition (2023.findings-acl)
Copied to clipboard
Zhen Zhang, Mengting Hu, Shiwan Zhao, Minlie Huang, Haotian Wang, Lemao Liu, Zhirui Zhang, Zhe Liu, Bingzhe Wu
| Challenge: | Named entity recognition (NER) systems focus on improving model performance, ignoring the need to quantify model uncertainty. |
| Approach: | They propose to introduce two uncertainty-guided loss terms to the conventional EDL and a series of uncertainty-guiding training strategies to solve these challenges. |
| Outcome: | The proposed method achieves better OOV/OOD detection performance and generalization ability on OOV entities compared to state-of-the-art methods. |
Applying BERT to Document Retrieval with Birch (D19-3)
Copied to clipboard
| Challenge: | Birch is an open-source document retrieval system that integrates with the Anserini information retrieval toolkit to demonstrate end-to-end search over large document collections. |
| Approach: | They propose to integrate Anserini with a BERT-based document ranking model that provides an end-to-end open-source search engine. |
| Outcome: | The proposed system outperforms existing approaches to document retrieval and question answering on standard newswire and social media test collections. |
Reasoning or Knowledge: Stratified Evaluation of Biomedical LLMs (2026.eacl-long)
Copied to clipboard
| Challenge: | Medical reasoning in large language models is a complex cognitive process through which clinicians interpret patient data and make diagnostic and therapeutic decisions. |
| Approach: | They propose an evaluation framework that disentangles knowledge recall from reasoning by training a PubMedBERT-based classifier and applying it to 11 widely used biomedical QA benchmarks. |
| Outcome: | The proposed evaluation framework disentangles knowledge recall from reasoning by training a PubMedBERT-based classifier and applying it to 11 widely used biomedical QA benchmarks. |
Recurrent Inference in Text Editing (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Existing inference methods map the unedited text to the edited text or to the editing operations, but performance is degraded by the limited source text encoding and long, varying decoding steps. |
| Approach: | They propose a new inference method that iteratively performs editing actions . they introduce three types of editing tasks: AOR, AES, AEC . |
| Outcome: | The proposed method significantly narrows the problem space by iterating editing actions. |
Triple-Hybrid Energy-based Model Makes Better Calibrated Natural Language Understanding Models (2023.eacl-main)
Copied to clipboard
| Challenge: | In-distribution (ID) miscalibration and out-of-difference (OOD) detection are main concerns for pre-trained language models. |
| Approach: | They propose a triple-hybrid EBM which combines the benefits of classifier, conditional generative model and marginal generative models altogether. |
| Outcome: | The proposed model outperforms previous methods in terms of ID calibration and OOD detection by a large margin while maintaining competitive accuracy. |
AgentCPM-GUI: Building Mobile-Use Agents with Reinforcement Fine-Tuning (2025.emnlp-demos)
Copied to clipboard
Zhong Zhang, Yaxi Lu, Yikun Fu, Yupeng Huo, Shenzhi Yang, Yesai Wu, Han Si, Xin Cong, Haotian Chen, Yankai Lin, Xie Xie, Wei Zhou, Wang Xu, Zhou Su, Zhongwu Zhai, Xiaoming Liu, null Meiyudong, Jianming Xu, Hongyan Tian, Chongyi Wang, Chi Chen, Yuan Yao, Zhiyuan Liu, Maosong Sun
| Challenge: | Large language model agents have enabled GUI-based automation, but their deployment is limited by noisy data, poor generalization, and lack of support for non-English GUIs. |
| Approach: | They propose an 8B-parameter GUI agent built for robust and efficient on-device GUI interaction. |
| Outcome: | The proposed GUI agent achieves promising performance on five public benchmarks and proposed Chinese benchmark CAGUI. |
D-Artemis: A Deliberative Cognitive Framework for Mobile GUI Multi-Agents (2026.findings-acl)
Copied to clipboard
Hongze Mi, Yibo Feng, WenJie Lu, Yuqi Wang, Jinyuan Li, Song Cao, He Cui, Tengfei Tian, Xuelin Zhang, Haotian Luo, Di Sun, Jun Fang, Hua Chai, Naiqiang Tan, Gang Pan
| Challenge: | Graphical User Interface (GUI) agents aim to automate a wide spectrum of human tasks by emulating user interaction. |
| Approach: | They propose a deliberative framework that leverages a fine-grained tip retrieval mechanism to inform its decision-making process. |
| Outcome: | The proposed framework achieves SOTA among open-source general models on AndroidWorld and ScreenSpot-V2 . it leverages a fine-grained, app-specific tip retrieval mechanism to inform its decision-making process . |
Reasoning under Uncertainty: Efficient LLM Inference via Unsupervised Confidence Dilution and Convergent Adaptive Sampling (2025.emnlp-main)
Copied to clipboard
Zhenning Shi, Yijia Zhu, Yi Xie, Junhan Shi, Guorui Xie, Haotian Zhang, Yong Jiang, Congcong Miao, Qing Li
| Challenge: | Large language models suffer from overconfidence and computational inefficiency due to fixed computation budgets and miscalibrated confidence estimates. |
| Approach: | They propose a framework for computationally efficient, trustworthy reasoning under uncertainty using Diversity-Aware Self-Signal Dilution and Convergent Adaptive Weighted Sampling techniques. |
| Outcome: | The proposed framework reduces inference cost by 70% while maintaining accuracy levels while reducing inference costs. |
GAP: a Global Adaptive Pruning Method for Large Language Models (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing structured pruning methods employ uniform compression rates across network layers, neglecting the varying importance of different network depths. |
| Approach: | They propose a pruning framework that minimizes global capability loss by layer-adaptive pruning rates. |
| Outcome: | The proposed approach achieves comparable performance with state-of-the-art methods at high pruning rates and shows significant advantages at low pruning rates. |
Thinking Economically: A Hierarchical Framework for Adaptive-Complexity Reasoning in LLMs (2026.findings-acl)
Copied to clipboard
Yubo Gao, Haotian Wu, Hong Chen, Junquan Huang, Yibo Yan, Jungang Li, Zihao Dongfang, Sicheng Tao, PS Tan, Jie Zhang, Xuming Hu
| Challenge: | Existing efficiency methods for Chain-of-Thought (CoT) generate excessively long rationales without commensurate accuracy gains. |
| Approach: | They propose a training framework that operationalizes this principle through coarse-to-fine budgeting. |
| Outcome: | Experiments on GSM8K and MATH500 show that HAB surpasses standard CoT in accuracy and reduces token usage, achieving stronger performance-efficiency trade-off than baselines. |
Cross-Domain Modeling of Sentence-Level Evidence for Document Retrieval (D19-1)
Copied to clipboard
| Challenge: | Existing test collections provide only document-level relevance judgments, and documents exceed the length that BERT was designed to handle. |
| Approach: | They propose to aggregate sentence-level evidence to rank news articles using BERT . they also leverage passage-level relevance judgments available in other domains to fine-tune BERT models that capture cross-domain notions of relevance. |
| Outcome: | The proposed model aggregates sentence-level evidence to rank documents on three standard test collections. |
UBench: Benchmarking Uncertainty in Large Language Models with Multiple Choice Questions (2025.findings-acl)
Copied to clipboard
Xunzhi Wang, Zhuowei Zhang, Gaonan Chen, Qiongyu Li, Bitong Luo, Zhixin Han, Haotian Wang, Zhiyu Li, Hang Gao, Mengting Hu
| Challenge: | Existing methods for benchmarking the uncertainty of large language models face challenges . existing methods require internal model access, additional training, or high computational costs . |
| Approach: | They propose a new benchmark for evaluating the uncertainty of large language models based on confidence intervals . UBench encompasses 11,978 multiple choice questions spanning knowledge, language, understanding, and reasoning capabilities. |
| Outcome: | The proposed method outperforms existing methods for benchmarking the uncertainty of large language models. |
OASIS: Order-Augmented Strategy for Improved Code Search (2025.acl-long)
Copied to clipboard
Gao Zuchen, Zizheng Zhan, Xianming Li, Erxin Yu, Haotian Zhang, Chenbin Chenbin, Yuqun Zhang, Jing Li
| Challenge: | Code embeddings capture the semantic representations of code and are crucial for various code-related large language model (LLM) applications. |
| Approach: | They propose an order-augmented strategy for improved code search that leverages order-based similarity labels to capture subtle differences in similarity among negative pairs. |
| Outcome: | The proposed model outperforms state-of-the-art models focusing on major positive-negative differences. |
Leveraging Multilingual Training for Authorship Representation: Enhancing Generalization across Languages and Domains (2025.emnlp-main)
Copied to clipboard
| Challenge: | Authorship representation (AR) models capture an author's distinctive writing style by encoding documents written by the same author as nearby vectors in the embedding space. |
| Approach: | They propose a method that incorporates probabilistic content masking and language-aware batching to improve contrastive learning by reducing cross-lingual interference. |
| Outcome: | The proposed model outperforms monolingual baselines in 21 out of 22 non-English languages and reaches a maximum gain of 15.91% in a single language. |
Towards Generating Controllable and Solvable Geometry Problem by Leveraging Symbolic Deduction Engine (2025.acl-industry)
Copied to clipboard
Zhuoxuan Jiang, Tianyang Zhang, Peiyan Peng, Jing Chen, Yinong Xun, Haotian Zhang, Lichi Li, Yong Li, Shaohua Zhang
| Challenge: | Compared to math word problems, geometry problems emphasize multi-modal formats and the translation between informal and formal languages. |
| Approach: | They propose a symbolic deduction engine-based geometry problem generation framework that leverages a symbolic deduction engine to generate geometry problems. |
| Outcome: | The proposed method avoids inherent biases in translating natural language into formal language and guarantees to control the generated problems in terms of knowledge points and difficulties by an elaborate checking function. |
Real or Robotic? Assessing Whether LLMs Accurately Simulate Qualities of Human Responses in Human-LLM Dialogue (2026.findings-acl)
Copied to clipboard
Jonathan Ivey, Shivani Kumar, Jiayu Liu, Hua Shen, Sushrita Rakshit, Rohan Raju, Haotian Zhang, Aparna Ananthasubramaniam, Junghwan Kim, Bowen Yi, Dustin Wright, Abraham Israeli, Anders Giovanni Møller, Lechen Zhang, David Jurgens
| Challenge: | Recent work has sought to use large language models to simulate human-human and human-LLM interactions. |
| Approach: | They use a large-scale dataset to generate a paired LLM-LLM and human-LLm dialogues from the WildChat dataset and quantify how well they align with their human counterparts. |
| Outcome: | The proposed models perform similarly in simulating English, Chinese, and Russian dialogues. |
VisFinEval: A Scenario-Driven Chinese Multimodal Benchmark for Holistic Financial Understanding (2025.emnlp-main)
Copied to clipboard
Zhaowei Liu, Xin Guo, Haotian Xia, Lingfeng Zeng, Fangqi Lou, Jinyi Niu, Mengping Li, Qi Qi, Jiahuan Li, Wei Zhang, Yinglong Wang, Weige Cai, Weining Shen, Liwen Zhang
| Challenge: | Existing benchmarks focus on text comprehension, but MLLMs lack the ability to integrate visual data over financial visuals. |
| Approach: | They evaluate 21 state-of-the-art multimodal large language models in a zero-shot setting . they use an annotated question–answer pair from eight common financial image modalities . |
| Outcome: | The new benchmark outperforms existing models but trailed financial experts by 14 percentage points. |
Grammar-Based Code Representation: Is It a Worthy Pursuit for LLMs? (2025.findings-acl)
Copied to clipboard
Qingyuan Liang, Zhao Zhang, Zeyu Sun, Zheng Lin, Qi Luo, Xiao Yueyi, Yizhou Chen, Yuqun Zhang, Haotian Zhang, Lu Zhang, Chenbin Chenbin, Yingfei Xiong
| Challenge: | Existing research demonstrates the effectiveness of grammar-based code representations in small-scale models, showing their ability to reduce syntax errors and enhance performance. |
| Approach: | They develop a series of billion-scale grammar-based code representations that incorporate grammar rules into the code generation process. |
| Outcome: | Experiments on HumanEval and MBPP show that grammar-based representations reduce syntax errors and improve performance even in billion-scale models. |
Crossroads, Buildings and Neighborhoods: A Dataset for Fine-grained Location Recognition (2022.naacl-main)
Copied to clipboard
| Challenge: | Named Entity Recognition (NER) datasets annotate coarse-grained entities such as a continent, a country, or a city. |
| Approach: | They propose a dataset HarveyNER with fine-grained locations annotated in tweets that characterizes many complex and long location mentions in informal descriptions. |
| Outcome: | The proposed dataset outperforms existing systems on hard cases and improves on the heuristic curricula. |
Natural Language Processing in Support of Evidence-based Medicine: A Scoping Review (2025.findings-acl)
Copied to clipboard
| Challenge: | Evidence-based medicine (EBM) is at the forefront of modern healthcare, emphasizing the use of the best available scientific evidence to guide clinical decisions. |
| Approach: | They propose to investigate the use of Natural Language Processing (NLP) techniques to identify, appraise, synthesize, apply, and disseminate evidence in EBM. |
| Outcome: | The proposed methods support the five fundamental steps of EBM—Ask, Acquire, Appraise, Apply, and Assess. |
FinEval: A Chinese Financial Domain Knowledge Evaluation Benchmark for Large Language Models (2025.naacl-long)
Copied to clipboard
Xin Guo, Haotian Xia, Zhaowei Liu, Hanyang Cao, Zhi Yang, Zhiqiang Liu, Sizhe Wang, Jinyi Niu, Chuqi Wang, Yanhui Wang, Xiaolong Liang, Xiaoming Huang, Bing Zhu, Zhongyu Wei, Yun Chen, Weining Shen, Liwen Zhang
| Challenge: | Large language models have demonstrated outstanding performance in various natural language processing tasks, but their security capabilities in the financial domain have not been explored. |
| Approach: | They propose to use a benchmark to evaluate large language models' financial domain knowledge and practical abilities. |
| Outcome: | The proposed benchmark evaluates large language models' financial domain knowledge and practical abilities. |
Causally Modeling the Linguistic and Social Factors that Predict Email Response (2025.naacl-long)
Copied to clipboard
Yinuo Xu, Hong Chen, Sushrita Rakshit, Aparna Ananthasubramaniam, Omkar Yadav, Mingqian Zheng, Michael Jiang, Lechen Zhang, Bowen Yi, Kenan Alkiek, Abraham Israeli, Bangzhao Shu, Hua Shen, Jiaxin Pei, Haotian Zhang, Miriam Schirmer, David Jurgens
| Challenge: | a key intent behind many emails is to get a reply from the recipient. |
| Approach: | They propose to model the intents, expectations, and responsiveness in email exchanges by using a dataset containing 1800 emails annotated with nuanced types of intents and expectations. |
| Outcome: | The proposed model is based on 1800 emails annotated with nuanced types of intents and expectations . it shows that social status, argumentation, and strength of social connection influence email response rates . |
DavIR: Data Selection via Implicit Reward for Large Language Models (2025.acl-long)
Copied to clipboard
Haotian Zhou, Tingkai Liu, Qianli Ma, Yufeng Zhang, Jianbo Yuan, Pengfei Liu, Yang You, Hongxia Yang
| Challenge: | 6% of Alpaca dataset selected with DavIR can steer both LLaMA and Gemma models to produce superior performance compared to the same models trained on the full 52K dataset. |
| Approach: | They propose a model-based data selection method for post-training Large Language Models . they generalize Reducible Holdout Loss to core-set selection problem of causal language modeling . |
| Outcome: | The proposed method can steer both LLaMA and Gemma models to superior performance compared to the same models trained on the full 52K dataset. |
CoAlign: Uncertainty Calibration of LLM for Geospatial Repartition (2025.acl-industry)
Copied to clipboard
| Challenge: | Existing methods to optimize geospatial repartition rely on manual adjustments by experts or algorithmic optimization using limited offline operational metrics. |
| Approach: | They propose a framework that calibrates LLM uncertainty to enable robust geospatial repartition by integrating historical data with LLM-generated candidates. |
| Outcome: | The proposed framework calibrates LLM uncertainty to enable robust geospatial repartition by integrating historical data with LLM-generated candidates. |
CORES: Code-Oriented Reasoning for Complex Text-to-SQL and Generalizable TableQA (2026.findings-acl)
Copied to clipboard
Meng Zhang, Ruochun Jin, Yuanxi Peng, Wenjing Yang, Haotian Wang, Liting Sun, Kun Hu, Silin Yang, Zhang Ke-di
| Challenge: | Text-to-SQL models struggle with complex analytical tasks such as generating simple SQL queries. |
| Approach: | They propose a text-to-sql model that leverages Python as a procedural reasoning pivot to enhance both complex SQL generation and tabular reasoning. |
| Outcome: | The proposed model outperforms baseline models on six text-to-SQL benchmarks by 6.44% on average while maintaining good capability on three tableQA benchmarks. |