Papers by Yuwei Wu
Beyond Token Length: Step Pruner for Efficient and Accurate Reasoning in Large Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing reinforcement learning methods for large reasoning models suffer from excessive verbosity, known as "overthinking." Existing models penalize generated tokens to promote conciseness, but these methods encounter two challenges: they may develop hacking behavior in later stages of training by discarding reasoning steps. |
| Approach: | They propose a framework that steers large reasoning models toward more efficient reasoning . they prioritize correctness while imposing penalties for redundant steps . |
| Outcome: | The proposed framework reduces token usage by 69.7% on AIME24. |
ZoFia: Zero-Shot Fake News Detection with Entity-Guided Retrieval and Multi-LLM Interaction (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) are limited by knowledge cutoff and can generate factual hallucinations when handling time-sensitive news. |
| Approach: | They propose a two-stage zero-shot fake news detection framework that uses a hierarchical salience and saliency-calibrated minimum margin of relevance algorithm to extract core entities accurately. |
| Outcome: | The proposed framework outperforms existing zero-shot baselines and even most few-shot methods on two public datasets. |
Personalized Response Generation via Generative Split Memory Network (2021.naacl-main)
Copied to clipboard
| Challenge: | Despite the success of text generation and dialogue systems, how to endow a text generation system with personality traits remains under-investigated. |
| Approach: | They propose a model to generate personalized responses on reddit using user profiles and posting histories. |
| Outcome: | The proposed model improves over the state-of-the-art response generation models. |
A Slot Is Not Built in One Utterance: Spoken Language Dialogs with Sub-Slots (2022.findings-acl)
Copied to clipboard
| Challenge: | Sub-Slot based task-oriented dialogs provide slot values segment by segment over multiple turns. |
| Approach: | They define a task called Sub-Slot based Task-Oriented Dialog (SSTOD) they build a Chinese dialog dataset SSD for boosting research on SSTOD. |
| Outcome: | The proposed task is called Sub-Slot based Task-Oriented Dialog (SSTOD) it includes 40K dialogs and 500K utterances from Chinese names, phone numbers, ID numbers and license plate numbers . the dataset is well annotated with sub-slot values, slot values, dialog states and actions . |
Chain-of-History Reasoning for Temporal Knowledge Graph Forecasting (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing graph-based models excel at capturing structural information within TKGs but lack semantic comprehension abilities. |
| Approach: | They propose a plug-and-play module to enhance the performance of graph-based TKG models by exploring high-order histories step-by-step. |
| Outcome: | Experiments on three datasets and backbones show that CoH is effective in capturing high-order historical information for LLMs. |
New Intent Discovery with Pre-training and Contrastive Learning (2022.acl-long)
Copied to clipboard
| Challenge: | Existing methods for identifying intents from unlabeled utterances are label-intensive, inefficient, and inaccurate. |
| Approach: | They propose a multi-task strategy to leverage unlabeled data and external labeled data for representation learning. |
| Outcome: | The proposed method outperforms state-of-the-art methods on three intent recognition benchmarks. |
Fine-tuning Pre-trained Language Models for Few-shot Intent Detection: Supervised Pre-training and Isotropization (2022.naacl-main)
Copied to clipboard
| Challenge: | Recent studies show that fine-tuning pre-trained language models with a small set of labeled utterances in a supervised manner is helpful, but it yields an anisotropic feature space, which may suppress the expressive power of the semantic representations. |
| Approach: | They propose to regularize supervised pre-training towards isotropy by contrastive learning and correlation matrix regularizers. |
| Outcome: | The proposed methods improve supervised pre-training by regularizing the feature space towards isotropy. |
From Scaffolding to Assimilation: Progressive Structural Internalization for Format-Constrained Creative Text Generation (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing paradigms rely on unreliable prompting or rigid constrained decoding strategies to achieve aesthetic unity. |
| Approach: | They propose a framework to embed external constraints into the model’s intrinsic intuition and use it to generate open-ended creative texts. |
| Outcome: | The proposed framework surpasses baselines in both strict constraint adherence and literary aesthetics. |
LLM as Prompter: Low-resource Inductive Reasoning on Arbitrary Knowledge Graphs (2024.findings-acl)
Copied to clipboard
| Challenge: | Knowledge Graph (KG) inductive reasoning is widely adopted in various applications. |
| Approach: | They propose a framework for low-resource inductive reasoning using Large Language Models to generate a graph-structural prompt for pre-trained KGs. |
| Outcome: | The proposed framework outperforms previous methods in three-shot, one-shot and zero-shot reasoning tasks. |
LPO: Towards Accurate GUI Agent Interaction via Location Preference Optimization (2026.findings-acl)
Copied to clipboard
Jiaqi Tang, Yu Xia, Yi-Feng Wu, Yuwei Hu, Chen Yuhui, Qing-Guo Chen, Xiaogang Xu, Xiangyu Wu, Hao LU, Yanqing Ma, Shiyin Lu, Qifeng Chen
| Challenge: | Existing strategies for spatial localization are limited due to their limited capacity to perceive positional data. |
| Approach: | They propose a location-based approach that leverages locational data to optimize interaction preferences. |
| Outcome: | The proposed approach achieves SOTA results across offline benchmarks and real-world evaluations. |
Learning Latent Relations for Temporal Knowledge Graph Reasoning (2023.acl-long)
Copied to clipboard
| Challenge: | Existing methods for Temporal Knowledge Graph reasoning capture intra- and inter-time latent relations between entities that appear at different times. |
| Approach: | They propose a Latent relations Learning method for TKG reasoning that captures latent relations between entities at different times. |
| Outcome: | The proposed method exploits the intra- and inter-time latent relations of entities at different times. |
MetaTKG: Learning Evolutionary Meta-Knowledge for Temporal Knowledge Graph Reasoning (2022.emnlp-main)
Copied to clipboard
| Challenge: | Existing models rely on historical information to learn embeddings for entities, but ignore the evolution of facts. |
| Approach: | They propose a Temporal Meta-learning framework to learn evolutionary meta-knowledge from TKGs. |
| Outcome: | The proposed method improves on four widely-used datasets and three backbones on a wide range of scenarios on tKGs. |
In-Context Compositional Generalization for Large Vision-Language Models (2024.emnlp-main)
Copied to clipboard
| Challenge: | Recent work shows that in-context learning for large language models exhibits compositional generalization capacity. |
| Approach: | They propose a method to exhibit in-context compositional generalization in large vision-language models by combining visual and linguistic modalities. |
| Outcome: | The proposed method reduces redundancy and complexity in in-context learning with LVLMs. |
MirrorCAPTCHA: Wild CAPTCHA, Wild Distribution, Wild Web-based Platform Meet Multimodal LLM Agents (2026.acl-long)
Copied to clipboard
Xiangyu Wu, Yuwei Hu, Tianyu Cui, Yueying Tian, Qing-Guo Chen, Zhao Xu, Weihua Luo, Kaifu Zhang, Yang Yang, Jianfeng Lu
| Challenge: | Existing agent benchmarks fail to evaluate an agent's real-world capacity to handle CAPTCHA . Existing benchmarks ignore this practical challenge, failing to evaluate agents' ability to handle complex visual CAPTchas. |
| Approach: | They propose a benchmark annotated with Weighted Pass Rate and a new metric to measure agent's ability to handle CAPTCHA. |
| Outcome: | The proposed benchmark outperforms current state-of-the-art closed-source models on mirrorCAPTCHA and achieves 9.4% higher average weighted pass rate and 2.13% higher average Completion degree. |
Effectiveness of Pre-training for Few-shot Intent Classification (2021.findings-emnlp)
Copied to clipboard
| Challenge: | Existing paradigms further pre-train language models such as BERT on vast amount of unlabeled corpus, but we find it highly effective and efficient to simply fine-tune BERT with roughly 1,000 labeled utterances from public datasets. |
| Approach: | They propose to fine-tune BERT with a small set of labeled utterances from public datasets to achieve a pre-trained model based on a set of 1,000 labeles. |
| Outcome: | The proposed model can outperform existing models on domains with very different semantics on novel domains. |