Papers by Yumeng Wang
KnowledgeBerg: Evaluating Systematic Knowledge Coverage and Compositional Reasoning in Large Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing LLMs lack systematic coverage of a bounded knowledge universe and compositional set-based reasoning over that universe. |
| Approach: | They propose a benchmark for multiple-choice questions based on 1,183 enumeration seeds . they use knowledge width, cardinality of required universe, reasoning depth to formalize the challenge . |
| Outcome: | The proposed benchmarks achieve only 5.26–36.88 F1 on universe enumeration and 16.00–44.19 accuracy on knowledge-grounded reasoning. |
CLEAN–EVAL: Clean Evaluation on Contaminated Large Language Models (2024.findings-naacl)
Copied to clipboard
Wenhong Zhu, Hongkun Hao, Zhiwei He, Yun-Ze Song, Jiao Yueyang, Yumeng Zhang, Hanxu Hu, Yiran Wei, Rui Wang, Hongyuan Lu
| Challenge: | Existing methods to evaluate large language models are prone to data contamination. |
| Approach: | They propose a method which parses contaminated data and back-translates it into a candidate set. |
| Outcome: | The proposed method reduces data contamination and evaluates the LLMs more cleanly. |
Static or Dynamic: Towards Query-Adaptive Token Selection for Video Question Answering (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches to compress video inputs ignore the importance of static and dynamic information in long videos, leading to inefficient token usage within limited budgets. |
| Approach: | They propose a token selection strategy that adaptively adjusts static and dynamic information based on question requirements. |
| Outcome: | The proposed method achieves performance improvements (up to 5.8%) on multiple video question answering benchmarks. |
Unveiling the Lack of LVLM Robustness to Fundamental Visual Variations: Why and Path Forward (2025.findings-acl)
Copied to clipboard
| Challenge: | Large Vision Language Models (LVLMs) have shown impressive performance on various vision-language tasks. |
| Approach: | They propose a benchmark framework for evaluating Visual Variation Robustness of Large Vision Language Models that incorporates automated evaluation dataset generation and principled metrics for thorough robustness assessment. |
| Outcome: | The proposed framework identifies a vulnerability to visual variations affecting even advanced models that excel at complex vision-language tasks but significantly underperform on simple tasks like object recognition. |
QUIDS: Query Intent Description for Exploratory Search via Dual Space Modeling (2025.emnlp-main)
Copied to clipboard
| Challenge: | Using QUIDS, we generate user-facing query intent descriptions that surface what the search engine likely inferred the query to mean based on post-retrieval evidence. |
| Approach: | They propose a method that leverages dual-space contrastive learning to isolate intent-relevant information while suppressing irrelevant content. |
| Outcome: | The proposed method outperforms state-of-the-art methods across ROUGE, BERTScore, and human/LLM evaluations. |
Mathematical Proof as a Litmus Test: Revealing Failure Modes of Advanced Large Reasoning Models (2026.acl-long)
Copied to clipboard
| Challenge: | Large reasoning models have demonstrated remarkable mathematical problem-solving abilities, but their true reasoning shortcomings are often hidden. |
| Approach: | They propose to leverage the rigor and methodological complexity of mathematical proofs as a diagnostic tool to expose hidden failures. |
| Outcome: | The proposed model evaluation exploits the rigor and complexity of proof problems to uncover 10 fine-grained errors. |
CALM: Unleashing the Cross-Lingual Self-Aligning Ability of Language Model Question Answering (2025.findings-naacl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) are pre-trained on extensive multilingual corpora to acquire both language-specific cultural knowledge and general knowledge. |
| Approach: | They propose to use the **C**ross-Lingual Self-**Aligning ability of **L**anguage **M**odels to align knowledge across languages. |
| Outcome: | The proposed model performs well in both zero-shot and retrieval-augmented settings. |
End-to-End Optimization for Multimodal Retrieval-Augmented Generation via Reward Backpropagation (2025.findings-emnlp)
Copied to clipboard
| Challenge: | MM-RAG is a promising approach for enhancing the reliability and factuality of large vision-language models . current methods focus on component-level optimizations and necessitate extensive component-specific training datasets . |
| Approach: | They propose a new paradigm that backpropagates global rewards to each component . this backpropage transforms local losses into specific local losses . |
| Outcome: | The proposed paradigm achieves high training efficiency on knowledge-intensive multimodal benchmarks. |
Embracing Imperfection: Simulating Students with Diverse Cognitive Levels Using LLM-based Agents (2025.acl-long)
Copied to clipboard
| Challenge: | Large language models (LLMs) are becoming increasingly popular in education, enabling researchers to simulate students' learning patterns and learning patterns. |
| Approach: | They propose a training-free framework for student simulation that takes into account student cognitive diversity and realism. |
| Outcome: | The proposed model outperforms baseline models and achieves 100% improvement in simulation accuracy and realism. |
LaERC-S: Improving LLM-based Emotion Recognition in Conversation with Speaker Characteristics (2025.coling-main)
Copied to clipboard
| Challenge: | Emotion recognition in conversation (ERC) is a task of discerning human emotions for each utterance within a conversation. |
| Approach: | They propose a framework that uses large language models to analyze speaker characteristics . they use two-stage learning to make the models reason speaker characteristics and track emotion of the speaker . |
| Outcome: | The proposed framework outperforms existing methods on three benchmark datasets. |