Papers by Hui Wei

35 papers
CodeArena: Evaluating and Aligning CodeLLMs on Human Preference (2025.emnlp-main)

Copied to clipboard

Challenge: Code large language models (codeLLMs) focus on synthesizing the correct code snippet, ignoring the alignment with human preferences.
Approach: They propose a benchmark code-based on 40 categories and 44 programming languages to emulate real-world coding tasks.
Outcome: The proposed benchmarks show that open-source code LLMs perform better than open-sourced ones.
IW-Bench: Evaluating Large Multimodal Models for Converting Image-to-Web (2025.findings-acl)

Copied to clipboard

Challenge: Existing models have been introduced to improve image comprehension, but there is no robust benchmark for imagetoweb conversion.
Approach: They propose a benchmark to assess imagetoweb conversion proficiency of large multimodal models . they propose to measure layout information of web pages by parsing the Document Object Model tree .
Outcome: The proposed benchmark measures the layout information of web pages—i.e., the positional relationships between elements—which has been overlooked by prior work.
Turning the Tide: Repository-based Code Reflection (2025.findings-emnlp)

Copied to clipboard

Challenge: Code large language models (LLMs) enhance programming by understanding and generating code across languages.
Approach: a new benchmark evaluates code understanding and generation in repositories using code large language models.
Outcome: The proposed model improves code understanding and generation in repositories by evaluating 1,888 test cases across 6 programming languages.
Temporal Scaling Law for Large Language Models (2025.emnlp-main)

Copied to clipboard

Challenge: Existing studies have found that the test loss of LLMs scales as power-laws with model size, computational budget, and dataset size.
Approach: They propose a concept of Temporal Scaling Law to study test loss of LLMs . they break down test loss into fine-grained token positions and develop a dynamic hyperbolic-law .
Outcome: The proposed model predicts the test loss of LLMs as the training steps scale up.
Do VLMs Have a Moral Backbone? A Study on the Fragile Morality of Vision-Language Models (2026.findings-acl)

Copied to clipboard

Challenge: Vision-Language Models (VLMs) have advanced multimodal learning, driving progress in cross-modal reasoning.
Approach: They propose to examine moral robustness of vision-language models by analyzing their moral stances under multimodal perturbations.
Outcome: The proposed model-agnostic multimodal perturbations expose VLMs to a variety of moral vulnerabilities, including a sycophancy trade-off where stronger instruction-following models are more susceptible to persuasion.
Towards Fine-Grained and Multi-Granular Contrastive Language-Speech Pre-training (2026.acl-long)

Copied to clipboard

Challenge: Existing models for fine-grained speaking styles are limited in terms of accuracy, coverage, and naturalness.
Approach: They propose a model that pre-trains with coarse captions and annotates with a pipeline that grounds captions in audio.
Outcome: The proposed model outperforms existing models with fine-grained style annotations . it integrates global and fine-granular supervision, enabling unified representations based on the proposed model .
VisiPruner: Decoding Discontinuous Cross-Modal Dynamics for Efficient Multimodal LLMs (2025.emnlp-main)

Copied to clipboard

Challenge: Multimodal Large Language Models (MLLMs) suffer from significant computational overhead due to the quadratic growth of attention computations with the number of multimodal tokens.
Approach: They propose a training-free pruning framework that prunes multimodal tokens without a trained pruning method.
Outcome: The proposed pruning framework outperforms existing token pruning methods and generalizes across diverse MLLMs.
TRELM: Towards Robust and Efficient Pre-training for Knowledge-Enhanced Language Models (2024.lrec-main)

Copied to clipboard

Challenge: Existing methods for incorporating external knowledge into language models do not prioritize learning embeddings for entity-related tokens.
Approach: They propose a framework for incorporating external knowledge into pre-training models that utilize entity-related tokens.
Outcome: The proposed framework reduces pre-training time by 50% and outperforms other KEPLMs in knowledge probing tasks and multiple knowledge-aware language understanding tasks.
PlanGenLLMs: A Modern Survey of LLM Planning Capabilities (2025.acl-long)

Copied to clipboard

Challenge: Existing studies have focused on developing LLMs to automate complex planning tasks.
Approach: They propose to provide a comprehensive overview of current LLM planners to fill this gap . they examine performance criteria including completeness, executability, optimality, representation, generalization, and efficiency .
Outcome: The proposed survey examines performance criteria for LLM planners and highlights their strengths and weaknesses.
MM-Align: Learning Optimal Transport-based Alignment Dynamics for Fast and Accurate Inference on Missing Modality Sequences (2022.emnlp-main)

Copied to clipboard

Challenge: Existing approaches to multimodal learning assume a complete input modality setting, i.e., each modality is either complete or completely missing in both training and test sets.
Approach: They propose an alignment dynamics learning module based on the theory of optimal transport for missing data imputation and a denoising training algorithm to enhance the quality of iputation and accuracy of model predictions.
Outcome: The proposed method performs faster and more accurate inferences under different missing conditions and alleviates the overfitting issue.
Anti-Length Shift: Dynamic Outlier Truncation for Training Efficient Reasoning Models (2026.acl-long)

Copied to clipboard

Challenge: Existing efficient reasoning methods rely on explicit length penalties for excessive verbosity on simple queries.
Approach: They propose a training-time intervention that selectively suppresses redundant tokens . they find length shift occurs when models generate unnecessary reasoning on trivial inputs - a phenomenon that is often unexplored .
Outcome: The proposed method reduces inference token usage by 78% while increasing accuracy compared to the initial policy and surpasses state-of-the-art efficient reasoning methods.
Facilitating Long Context Understanding via Supervised Chain-of-Thought Reasoning (2025.emnlp-main)

Copied to clipboard

Challenge: Recent advances in Large Language Models (LLMs) have enabled them to process increasingly longer sequences, ranging from 2K to 2M tokens and even beyond.
Approach: They propose a synthetic dataset in the financial domain that integrates Chain-of-Thought reasoning into LLMs in a supervised manner to facilitate effective long-context understanding.
Outcome: The proposed model outperforms standard GPT-4o-mini on the Loong benchmark and fine tunes LLaMA-3.1-8B-Instruct on the model, achieving a 28.0% gain on the financial subset.
Forget the Token and Pixel: Rethinking Gradient Ascent for Concept Unlearning in Multimodal Generative Models (2025.findings-acl)

Copied to clipboard

Challenge: Gradient Ascent (GA) has emerged as a promising approach for concept unlearning in Multimodal Generative Models (MGMs).
Approach: They propose a novel approach that selectively applies GA to targeted Conceptual Knowledge while preserving Natural Knowledge through Gradient Descent (GD).
Outcome: The proposed approach removes Conceptual Knowledge and inadvertently diminishes Natural Knowledge, resulting in utility degradation.
DoubleMix: Simple Interpolation-Based Data Augmentation for Text Classification (2022.coling-1)

Copied to clipboard

Challenge: Existing methods to improve the robustness of text classification models are token-, sentence-, and hiddenlevel augmentation.
Approach: They propose an interpolation-based data augmentation approach called DoubleMix to improve the robustness of text classification models by learning the “shifted” features in hidden space.
Outcome: The proposed approach outperforms several popular methods on six text classification benchmark datasets and visual analysis shows that the model features are highly interpretable.
SUN: Exploring Intrinsic Uncertainties in Text-to-SQL Parsers (2022.coling-1)

Copied to clipboard

Challenge: Existing methods that learn from multiple semantically-equivalent questions are limited to one-to-one mapping .
Approach: They propose a constraint to explore the underlying complementary semantic information among multiple semantically-equivalent questions and learn robust feature representations with reduced spurious associations.
Outcome: The proposed method outperforms strong competitors and achieves state-of-the-art results on five benchmark datasets.
Toward Safe and Human-Aligned Game Conversational Recommendation via Multi-Agent Decomposition (2026.findings-eacl)

Copied to clipboard

Challenge: Existing systems for conversational recommender systems (CRS) have strong results in movies, but games present distinct challenges . MATCHA framework provides specialized agents for intent parsing, tool-augmented retrieval, multi-LLM ranking, and stronger safety.
Approach: They propose a framework for conversational recommender systems that assigns specialized agents for intent parsing, tool-augmented retrieval, multi-LLM ranking and risk control.
Outcome: MATCHA outperforms baselines on real user request dataset, improves Hit@5 by 20%, reduces popularity bias by 24%, and achieves 97.9% adversarial defense.
PREMISE: Matching-based Prediction for Accurate Review Recommendation (2025.findings-naacl)

Copied to clipboard

Challenge: Experimental results show that PREMISE achieves promising performance with less computational cost.
Approach: They propose a new architecture for matching-based learning in multimodal fields for the MRHP task.
Outcome: The proposed architecture significantly boosts performance on multimodal tasks with less computational cost compared to the state-of-the-art fusion-based methods.
SVD-GCL: A Noise-Augmented Hybrid Graph Contrastive Learning Framework for Recommendation (2025.coling-main)

Copied to clipboard

Challenge: Recent advances in graph neural networks have made it difficult to capture user preferences.
Approach: They propose a graph contrastive learning recommendation model based on noise augmentation that integrates truncated singular value decomposition in the feature engineering stage.
Outcome: The proposed model reduces dimensionality and denoises the original data.
AIR: Complex Instruction Generation via Automatic Iterative Refinement (2025.emnlp-main)

Copied to clipboard

Challenge: Existing methods for generating complex instructions are resource-intensive and lack diversity.
Approach: They propose a framework to generate complex instructions with constraints using a document-generated initial instruction and an iterative refinement framework to incorporate LLM-as-judge guidance.
Outcome: The proposed framework significantly outperforms existing methods for generating complex instructions, and outperformed existing methods.
TokenSelect: Efficient Long-Context Inference and Length Extrapolation for LLMs via Dynamic Token-Level KV Cache Selection (2025.emnlp-main)

Copied to clipboard

Challenge: Rapid advances in Large Language Models have spurred demand for processing extended context sequences . however, performance degradation due to sequence lengths out-of-distribution and excessively long inference times are limiting LLMs in long-context scenarios.
Approach: They propose a training-free method for efficient and accurate long-context inference . they selectively involves a few critical KV cache tokens in attention calculation .
Outcome: The proposed method speeds up attention computation and accelerates inference time while reducing selection overhead.
Fast Quiet-STaR: Thinking Without Thought Tokens (2025.findings-emnlp)

Copied to clipboard

Challenge: Large Language Models have achieved impressive performance across a range of tasks, but further gains require more than scaling up model sizes or training data.
Approach: They propose a method that gradually reduces the number of thought tokens . this method allows models to internalize more abstract reasoning processes .
Outcome: The proposed framework preserves the benefits of token-level reasoning while reducing computational cost.
Improving Multimodal Fusion with Hierarchical Mutual Information Maximization for Multimodal Sentiment Analysis (2021.emnlp-main)

Copied to clipboard

Challenge: Existing work on multimodal sentiment analysis relies on back-propagated task loss or geometric property of feature spaces to produce favorable fusion results.
Approach: They propose a framework which hierarchically maximizes the Mutual Information (MI) in unimodal input pairs and between multimodal fusion result and unimod input to maintain task-related information through multimodal integration.
Outcome: The proposed framework maximizes the Mutual Information (MI) in unimodal input pairs and between multimodal fusion result and unimodulated input to maintain task-related information through multimodal integration.
SAT: Improving Semi-Supervised Text Classification with Simple Instance-Adaptive Self-Training (2022.findings-emnlp)

Copied to clipboard

Challenge: Existing methods for semi-supervised text classification have shown great performance in few-shot scenarios, where both labeled and unlabeled data are utilized.
Approach: They propose a simple instance-adaptive self-training method for semi-supervised text classification that generates two augmented views for each unlabeled data and trains a meta learner to identify relative strength of augmentations based on the similarity between the original view and the augmented view.
Outcome: The proposed method consistently shows competitive performance with varying sizes of labeled training data compared to existing semi-supervised learning methods.
UER: An Open-Source Toolkit for Pre-training Models (D19-3)

Copied to clipboard

Challenge: Existing work on pre-training models have shown that it is important to use a framework to deploy various pre- training models efficiently.
Approach: They propose an assemble-on-demand pre-training toolkit that assembles pre-trained models on demand and encapsulates them with rich modules.
Outcome: The proposed framework can reproduce state-of-the-art models or develop models that remain unexplored.
SANCL: Multimodal Review Helpfulness Prediction with Selective Attention and Natural Contrastive Learning (2022.coling-1)

Copied to clipboard

Challenge: e-commerce has become a research hotspot for review helpfulness prediction . a new approach to help predict helpfulness of multimodal product reviews is proposed .
Approach: They propose a machine learning task to identify helpfulness of multimodal product reviews . they use a probe-based strategy to enforce high attention weights on regions of greater significance .
Outcome: The proposed model achieves state-of-the-art performance with lower memory consumption on two benchmark datasets with three categories.
Breaking the Stage Barrier: A Novel Single-Stage Approach to Long Context Extension for Large Language Models (2025.coling-main)

Copied to clipboard

Challenge: Recent studies show that Large language models struggle with handling long token sequences due to limited training context size.
Approach: They propose a single-stage continual pretraining method to equip LLMs with long context modeling capabilities.
Outcome: The proposed method outperforms existing methods on 4 language modeling benchmarks.
Qwen2.5-xCoder: Multi-Agent Collaboration for Multilingual Code Instruction Tuning (2025.acl-long)

Copied to clipboard

Challenge: Existing methods to train code LLMs view each programming language in isolation . experimental results show that Qwen2.5-xCoder can bridge the gap between different programming languages .
Approach: They propose a framework that allows agents to collaborate to enhance multilingual instruction tuning for code LLMs.
Outcome: Experimental results show that Qwen2.5-xCoder can transfer knowledge efficiently and effectively between languages.
DSMoE: Matrix-Partitioned Experts with Dynamic Routing for Computation-Efficient Dense LLMs (2025.emnlp-main)

Copied to clipboard

Challenge: Existing sparsification methods like pruning can lose model knowledge through parameter removal.
Approach: They propose a novel approach that achieves sparsification by partitioning pre-trained FFN layers into computational blocks.
Outcome: The proposed approach achieves superior performance across language modeling and downstream tasks under equivalent computational constraints.
Learning to Instruct: Fine-Tuning a Task-Aware Instruction Optimizer for Black-Box LLMs (2025.findings-emnlp)

Copied to clipboard

Challenge: Learning to Instruct is a new paradigm for black-box LLMs with inaccessible internal states.
Approach: They propose a new paradigm that formulates instruction optimization as an LLM fine-tuning objective for a white-box “instruction engineer” LLM.
Outcome: The proposed framework outperforms strong baselines in performance and efficiency.
Monte Carlo Tree Search Based Prompt Autogeneration for Jailbreak Attacks against LLMs (2025.coling-main)

Copied to clipboard

Challenge: Jailbreak attacks craft specific prompts or append adversarial suffixes to prompts, thereby inducing language models to generate harmful or unethical content and bypassing the model’s safety guardrails.
Approach: They propose a Monte Carlo Tree Search (MCTS) based Prompt Auto-generation (MPA) method to generate adversarial suffixes for valid jailbreak attacks.
Outcome: The proposed method outperforms existing methods on open-source and closed-source models and shows that it can generate harmful responses.
Enhancing Multimodal Named Entity Recognition through Adaptive Mixup Image Augmentation (2025.coling-main)

Copied to clipboard

Challenge: Current named entity recognition methods struggle with text-image mismatch problem due to a lack of visual context.
Approach: They propose an adaptive mixup image augmentation method that generates augmented images based on matching score between text and image .
Outcome: The proposed method can be integrated into existing models and demonstrate consistent performance improvements.
MiLe Loss: a New Loss for Mitigating the Bias of Learning Difficulties in Generative Language Models (2024.findings-naacl)

Copied to clipboard

Challenge: Existing generative language models neglect an inherent challenge in text corpus during training, i.e., the imbalance between frequent tokens and infrequent ones.
Approach: They propose a function to mitigate the imbalance between frequent and infrequent tokens . authors propose 'MiLe Loss' function to assess learning difficulty of tokens during training .
Outcome: Experiments show that models with proposed model can improve on downstream benchmarks.
AgentMark: Utility-Preserving Behavioral Watermarking for Agents (2026.acl-long)

Copied to clipboard

Challenge: Recent advances in large language models (LLMs) have improved text generation and reasoning.
Approach: They propose a behavioral watermarking framework that embeds multi-bit identifiers into planning decisions while preserving utility.
Outcome: The proposed framework embeds multi-bit provenance into planning decisions while preserving utility.
Ranking-Based Autoencoder for Extreme Multi-label Classification (N19-1)

Copied to clipboard

Challenge: Existing methods to solve label dependency and noisy labeling problems are limited . experimental results show the proposed method is competitive to state-of-the-art methods .
Approach: They propose a deep learning XML method with word-vector-based self-attention followed by ranking-based AutoEncoder architecture to solve these problems.
Outcome: The proposed method is competitive to state-of-the-art methods on benchmark datasets.
Self-Adaptive Sampling for Accurate Video Question Answering on Image Text Models (2024.findings-naacl)

Copied to clipboard

Challenge: Image–text models (ITMs) are the prevalent architecture to solve video question–answering tasks, which requires only a few input frames to save huge computational cost compared to video–language models.
Approach: They propose a sampling method based on question–frame correlation that is efficient for the few-frame situations.
Outcome: The proposed method can boost the performance of image–text pretrained models and have a wide application scenario in terms of model architectures and dataset types.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations