Papers by Haoxuan Li
What’s Missing in Vision-Language Models? Probing Their Struggles with Causal Order Reasoning (2026.eacl-long)
Copied to clipboard
| Challenge: | Existing benchmarks often include a mixture of reasoning questions, making it difficult to truly assess VLMs’ causal reasoning abilities. |
| Approach: | They propose two new benchmarks specifically designed to isolate and rigorously evaluate VLMs’ causal reasoning abilities. |
| Outcome: | The proposed benchmarks show that vision-language models perform poorly on causal reasoning tasks, often only marginally surpassing random guessing. |
CharacterBox: Evaluating the Role-Playing Capabilities of LLMs in Text-Based Virtual Worlds (2025.naacl-long)
Copied to clipboard
| Challenge: | Evaluating role-playing capabilities in large language models is challenging due to complex dynamics involved in role-playering. |
| Approach: | They propose a simulation sandbox that generates situational fine-grained character behavior trajectories to enhance LLM performance. |
| Outcome: | The proposed model generates situational fine-grained character behavior trajectories to enhance performance. |
Text Embedding as Treatment: A Meta Causal Approach for Robust Sentiment Classification (2026.findings-acl)
Copied to clipboard
Fengxiang Cheng, Chuan Zhou, Xiang Li, Haoxuan Li, Wen-li Wang, Jinkun Chen, Mingming Gong, Kun Zhang
| Challenge: | Existing methods for sentiment classification use binary treatment of words . Existing approaches limit generalizability to novel words and low-frequency words if there is a word in a sentence that is not treated . |
| Approach: | They propose a meta-causal approach that uses a single training task to identify causal words for arbitrary words. |
| Outcome: | The proposed method reduces the spurious correlation between word treatment and sentiment classification by removing words with low treatment effects from a pre-trained language model. |
Beyond Function-Level Search: Repository-Aware Dual-Encoder Code Retrieval with Adversarial Verification (2025.findings-emnlp)
Copied to clipboard
| Challenge: | escalating complexity of modern codebases has intensified the need for code retrieval systems capable of interpreting cross-component change intents. |
| Approach: | RepoAlignBench is a benchmark designed to evaluate repository-level code retrieval . the benchmark proposes an adversarial reflection-augmented dual-tower architecture . |
| Outcome: | The proposed framework achieves 12.2% Top-5 Accuracy and 7.1% Recall improvements over state-of-the-art benchmarks. |
MT-Video-Bench: A Holistic Video Understanding Benchmark for Evaluating Multimodal LLMs in Multi-Turn Dialogues (2026.findings-acl)
Copied to clipboard
Yaning Pan, Qianqian Xie, Guohui Zhang, Zekun Moore Wang, Yongqian Wen, Yuanxing Zhang, Haoxuan Hu, Zhiyu Pan, Yibing Huang, Zhidong Gan, Yonghong Lin, An Ping, Shihao Li, Yanghai Wang, Tianhao Peng, Jiaheng Liu
| Challenge: | Existing evaluation benchmarks for Multimodal Large Language Models (MLLMs) focus on single-turn question answering, overlooking the complexity of multi-turn dialogues in real-world scenarios. |
| Approach: | They propose a video understanding benchmark for MLLMs in multi-turn dialogues that assesses six core competencies that focus on perceptivity and interactivity. |
| Outcome: | The MT-Video-Bench evaluates 1,000 multi-turn dialogues from diverse domains and reveals significant performance discrepancies and limitations in handling multi-turned video dialogues. |
Generative Music Models’ Alignment with Professional and Amateur Users’ Expectations (2025.findings-acl)
Copied to clipboard
| Challenge: | Recent years have witnessed rapid advances in text-to-music generation using large language models. |
| Approach: | They propose a task to align AI-generated music with human expressions . they use a dataset of over 1.5 million songs to analyze their content . |
| Outcome: | The proposed framework outperforms baseline models and facilitates end-to-end generation of songs audio. |
SimPBL: A Multi-Agent Framework for Project-Based Learning (2026.acl-long)
Copied to clipboard
Daniel Zhang-Li, Joy Jia Yin Lim, Binglin Liu, Shangqing Tu, Zijun Yao, Hao Peng, Jifan Yu, Haoxuan Li, Zhanxin Hao, Ye He, Zekun Li, Jiangyi Wang, Lei Hou, Bin Xu, Xin Cong, Zhiyuan Liu, Huiqin Liu, Yu Zhang, Juanzi Li
| Challenge: | Existing LLMs provide partial assistance without modeling these roles, and overly comprehensive help can reduce learner autonomy. |
| Approach: | They propose a multi-agent framework with an orchestrator agent that provides adaptive scaffolding from interaction logs and collaborator agents that support project work through boundary-aware collaboration. |
| Outcome: | The proposed framework improves learner examination scores by 14% . it is based on a multi-agent framework with an orchestrator agent . |
Unsupervised Vision-and-Language Pre-training Without Parallel Images and Captions (2021.naacl-main)
Copied to clipboard
| Challenge: | Existing models require large amounts of image-caption data for pre-training . existing models require expensive data collection and curation . |
| Approach: | They propose to conduct "mask-and-predict" pre-training on text-only and image-only corpora and introduce the object tags detected by an object recognition model as anchor points to bridge two modalities. |
| Outcome: | The proposed approach achieves performance close to a model pre-trained with aligned data, on four English benchmarks. |
Beyond Self-Report: Bridging the Intention-Behavior Gap in Critical Thinking Assessment via Interpretable Multi-Agent System (2026.acl-long)
Copied to clipboard
Zekun Li, Jifan Yu, Haoxuan Li, Ye He, Daniel Zhang-Li, Shangqing Tu, Joy Jia Yin Lim, Yikun Jiang, Jiaxin Yuan, Yu Zhang
| Challenge: | Accurate assessment of critical thinking is limited by the Intention Behavior Gap in psychology . evaluators that measure self-reported competence are limited by multiagent architectures . |
| Approach: | They propose a framework that operationalizes cognitive assessment into an interpretable multi-agent workflow with Assessment Chain-of-Thought. |
| Outcome: | The proposed framework aligns better with human expert ratings than gold-standard inventories on large-scale simulations and human participants. |
IGenBench: Benchmarking the Reliability of Text-to-Infographic Generation (2026.acl-long)
Copied to clipboard
Yinghao Tang, Xueding Liu, Boyuan Zhang, Tingfeng Lan, Yupeng Xie, Jiale Lao, Yiyao Wang, Haoxuan Li, Tingting Gao, Bo Pan, Luoxuan Weng, Xiuqi Huang, Minfeng Zhu, Yingchaojie Feng, Yuyu Luo, Wei Chen
| Challenge: | Generated infographics may appear correct at first glance but contain easily overlooked issues, such as distorted data encoding or incorrect textual content. |
| Approach: | They propose to evaluate reliability of text-to-infographic generation using IGenBench . they employ multimodal large language models to verify each question . |
| Outcome: | The proposed framework decomposes reliability verification into atomic yes/no questions based on a taxonomy of 10 question types. |
Graph of Trace: Visualizing Execution Traces of Scientific Agents (2026.acl-demo)
Copied to clipboard
| Challenge: | Scientific AI agents can perform complex research tasks, but these unfolded workflows are difficult for humans to inspect and review, limiting interpretable, controllable and effective human–AI collaboration. |
| Approach: | They propose a monitoring and visualization framework that records fine-grained execution events and organizes them into a directed graph that makes agent workflows explicit as they proceed. |
| Outcome: | The proposed framework records intermediate steps (e.g. tool calls and code executions) and renders them as real-time updated visual traces that expose workflow structure. |
Mitigating Spurious Correlations via Counterfactual Contrastive Learning (2025.findings-emnlp)
Copied to clipboard
Fengxiang Cheng, Chuan Zhou, Xiang Li, Alina Leidinger, Haoxuan Li, Mingming Gong, Fenrong Liu, Robert Van Rooij
| Challenge: | Existing methods to distinguish causally related words from spurious correlations are limited by the number of causally correlated words in a sentence. |
| Approach: | They propose to use probabilistic probability of necessity and probability of sufficiency to identify causal relationships rather than spurious correlations between words and class labels. |
| Outcome: | The proposed method is based on a contrastive learning approach name CPNS and is validated on public datasets. |
Understanding ME? Multimodal Evaluation for Fine-grained Visual Commonsense (2022.emnlp-main)
Copied to clipboard
| Challenge: | Existing models that understand image and text but also cross-reference in-between are lacking in evaluation data resources. |
| Approach: | They propose a multimodal evaluation pipeline to automatically generate question-answer pairs to test models’ understanding of the visual scene, text, and related knowledge. |
| Outcome: | The proposed model can answer the highly semantic VCR question correctly but fails to answer related visual question (Q2), textual question (q3), and background knowledge question ( Q4) as shallow mappings with language priors and unbalanced utilization of information between modalities. |
From Knowing to Teaching: Scaffolding Pedagogical Decisions for LLM Agent (2026.acl-long)
Copied to clipboard
Yucheng Wang, Shen Yang, Jifan Yu, Haoxuan Li, Joy Jia Yin Lim, Daniel Zhang-Li, Huiqin Liu, Lei Hou, Juanzi Li, Bin Xu
| Challenge: | Large language models produce content lacking pedagogical depth when asked to generate lessons . |
| Approach: | They propose a framework that allows teachers to select content according to pedagogical intent and sequence topics so foundations precede applications. |
| Outcome: | The framework achieves 67.8% win rate in human evaluation and 79.6% in LLM-based evaluation against eight baselines. |
Dataset Bias Mitigation in Multiple-Choice Visual Question Answering and Beyond (2023.findings-emnlp)
Copied to clipboard
Zhecan Wang, Long Chen, Haoxuan You, Keyang Xu, Yicheng He, Wenhao Li, Noel Codella, Kai-Wei Chang, Shih-Fu Chang
| Challenge: | Existing studies have examined dataset biases in VQA benchmarks with short-phrase answers Multiple-choice Question with the LONG Answers (VCR, VLEP, etc.) |
| Approach: | They propose to use Adversarial Data Synthesis (ADS) to generate synthetic training and debiased evaluation data and introduce Intra-sample Counterfactual Training (ICT) to assist models in utilizing synthesized training data. |
| Outcome: | The proposed approach improves model performance even in domain-shifted scenarios. |