Papers by Yuyu Zhang
Data Interpreter: An LLM Agent for Data Science (2025.findings-acl)
Copied to clipboard
Sirui Hong, Yizhang Lin, Bang Liu, Bangbang Liu, Binhao Wu, Ceyao Zhang, Danyang Li, Jiaqi Chen, Jiayi Zhang, Jinlin Wang, Li Zhang, Lingyao Zhang, Min Yang, Mingchen Zhuge, Taicheng Guo, Tuo Zhou, Wei Tao, Robert Tang, Xiangtao Lu, Xiawu Zheng, Xinbing Liang, Yaying Fei, Yuheng Cheng, Yongxin Ni, Zhibin Gou, Zongze Xu, Yuyu Luo, Chenglin Wu
| Challenge: | Large Language Models (LLMs) excel in various domains but face challenges when applied to data science workflows due to their complex, multi-stage nature. |
| Approach: | They propose a hierarchical graph-based agent that represents complexity and a progressive strategy for step-by-step verification, refinement, and consistent context management. |
| Outcome: | The proposed agent surpasses state-of-the-art baselines on the MATH dataset and performs better on InfiAgent-DABench. |
Self-Supervised Prompt Optimization (2025.findings-emnlp)
Copied to clipboard
Jinyu Xiang, Jiayi Zhang, Zhaoyang Yu, Xinbing Liang, Fengwei Teng, Jinhao Tu, Fashen Ren, Xiangru Tang, Sirui Hong, Chenglin Wu, Yuyu Luo
| Challenge: | Existing prompt optimization methods rely heavily on external references such as ground truth or by humans, limiting their applicability in real-world scenarios where such data is unavailable or costly to obtain. |
| Approach: | They propose a cost-efficient framework that discovers effective prompts for both closed and open-ended tasks without external reference. |
| Outcome: | The proposed framework outperforms state-of-the-art prompt optimization methods with significantly lower costs and fewer samples. |
SWE-QA-Pro: A Representative Benchmark and Scalable Training Recipe for Repository-Level Code Understanding (2026.findings-acl)
Copied to clipboard
Songcheng Cai, Zhiheng Lyu, Yuansheng Ni, Xiangchao Chen, Baichuan Zhou, Shenzhe Zhu, Yi Lu, Haozhe Wang, Chi Ruan, Benjamin Schneider, Weixu Zhang, Xiang Li, Andy Zheng, Yuyu Zhang, Ping Nie, Wenhu Chen
| Challenge: | Existing benchmarks for agentic repository-level code understanding overlook long tail topics and rely on memorized knowledge. |
| Approach: | They propose a repository-level agentic code understanding benchmark that uses long-tail repositories with executable environments to enforce topical balance. |
| Outcome: | Empirically, a Qwen3-8B model trained with the proposed benchmark outperforms GPT-4o by 2.3 points. |
GPT-Fathom: Benchmarking Large Language Models to Decipher the Evolutionary Path towards GPT-4 and Beyond (2024.findings-naacl)
Copied to clipboard
| Challenge: | Existing LLM leaderboards often reference scores reported in other papers without consistent settings and prompts, which may encourage cherry-picking favored settings and for better results. |
| Approach: | They propose an open-source and reproducible LLM evaluation suite built on top of OpenAI Evals that systematically evaluates 10+ leading LLMs and OpenAI’s legacy models on 20+ curated benchmarks across 7 capability categories. |
| Outcome: | The evaluation suite is built on top of OpenAI Evals and evaluates 10+ leading LLMs and OpenAI’s legacy models on 20+ curated benchmarks across 7 capability categories. |
Language Modeling with Shared Grammar (P19-1)
Copied to clipboard
| Challenge: | Recent work on structure-aware models have shown promising results on language modeling, but how to incorporate structure knowledge on corpus without syntactic annotations remains an open problem. |
| Approach: | They propose a neural variational language model which enables the sharing of grammar knowledge among different corpora. |
| Outcome: | The proposed model converges significantly faster to lower perplexity on two popular benchmark datasets. |
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. |
Question Directed Graph Attention Network for Numerical Reasoning over Text (2020.emnlp-main)
Copied to clipboard
Kunlong Chen, Weidi Xu, Xingyi Cheng, Zou Xiaochuan, Yuyu Zhang, Le Song, Taifeng Wang, Yuan Qi, Wei Chu
| Challenge: | Numerical reasoning requires both natural language understanding and arithmetic computation. |
| Approach: | They propose a graph representation for the context of the passage and question needed for numerical reasoning. |
| Outcome: | The proposed model achieves remarkable results in benchmark datasets such as DROP. |
MEBench: Benchmarking Large Language Models for Cross-Document Multi-Entity Question Answering (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) and Retrieval-augmented Generation (RAG) systems show promise, but their performance on cross-document MEQA remains underexplored due to the lack of tailored benchmarks. |
| Approach: | They propose a scalable multi-document, multi-entity benchmark to evaluate LLMs' capacity to retrieve, consolidate, and reason over scattered and dense information. |
| Outcome: | The proposed benchmarks show that even advanced models achieve only 59% accuracy on MEBench. |
MAR: Matching-Augmented Reasoning for Enhancing Visual-based Entity Question Answering (2024.emnlp-main)
Copied to clipboard
| Challenge: | Multimodal large language models (MLLMs) struggle with visual-based entity questions (VEQA) MLLM can identify A, but may refrain from answering due to privacy concerns. |
| Approach: | They propose a method that uses vector representations to analyze visual-based entity questions (VEQA) they use visual cues and textual information to integrate visual cus and visual information . |
| Outcome: | The proposed method significantly improves visual-based entity question answering (VEQA) it can identify faces, names, and alignments within visual objects, and then derive the answer over this matching graph. |
SciImpact: A Multi-Dimensional, Multi-Field Benchmark for Scientific Impact Prediction (2026.findings-acl)
Copied to clipboard
| Challenge: | Prior work on scientific impact prediction has focused on citation counts and its variants, leaving limited evaluation of models’ capability to reason about other dimensions. |
| Approach: | They propose a large-scale, multi-dimensional benchmark for scientific impact prediction spanning 19 fields. |
| Outcome: | The proposed model outperforms larger models and close-source models in a wide range of fields and measures of scientific impact across 19 fields. |
Beyond Single-shot Writing: Deep Research Agents are Unreliable at Multi-turn Report Revision (2026.acl-long)
Copied to clipboard
| Challenge: | Existing benchmarks for Deep Research Agents (DRAs) treat report generation as a single-shot writing task. |
| Approach: | They propose an evaluation suite that establishes multi-turn report revision as a new axis. |
| Outcome: | The evaluation suite establishes multi-turn report revision as a new axis. |
Concise Math Reasoning via Difficulty-Aware Distillation (2026.findings-acl)
Copied to clipboard
Yifan Wu, Jingze Shi, Bingheng Wu, Jiayi Zhang, Xiaotian Lin, Yizhang Zhu, Zhaoyang Yu, Bang Liu, Chenglin Wu, Nan Tang, Yuyu Luo
| Challenge: | Human experts tackle difficult math problems by identifying and executing a few pivotal steps rather than listing every intermediate thought. |
| Approach: | They propose a method for producing training data that mirrors concise human reasoning by rewriting a problem's solution to retain only the essential steps. |
| Outcome: | The proposed method outperforms models trained on 800k long CoT and cuts training and inference costs. |