Papers by Weijia Li

13 papers
Stop Looking for “Important Tokens” in Multimodal Language Models: Duplication Matters More (2025.emnlp-main)

Copied to clipboard

Challenge: Vision tokens in multimodal large language models often dominate computational overhead due to excessive length compared to linguistic modality.
Approach: They propose a token pruning method which defines an importance criterion for vision tokens and prunes the unimportant vision token during inference.
Outcome: The proposed method can prune 88.9% of vision tokens while maintaining comparable performance.
Token Pruning in Multimodal Large Language Models: Are We Solving the Right Problem? (2025.findings-acl)

Copied to clipboard

Challenge: Multimodal large language models have shown remarkable performance for cross-modal understanding and generation, yet suffer from severe inference costs.
Approach: They propose to prune redundant tokens in MLLMs to reduce computation and storage costs.
Outcome: The proposed method reduces the computational and storage costs of MLLMs by identifying redundant tokens and pruning them.
Beyond Surface-Level Pattern Trap: LLM Agents for Faster and Smarter Cross-Architecture Code Migration (2026.findings-acl)

Copied to clipboard

Challenge: cross-architecture code migration is a resource-intensive and errorprone task.
Approach: a framework for cross-architecture code migration is proposed to decouple implementation details through functional mining and code refactoring.
Outcome: a new framework improves performance and correctness over state-of-the-art frameworks on OpenCV migration tasks.
RoZO: Geometry-Aware Zeroth-Order Fine-Tuning on Low-Rank Adapters for Black-Box Large Language Models (2026.eacl-long)

Copied to clipboard

Challenge: Large language models (LLMs) have demonstrated exceptional performance across a wide range of tasks, yet fine-tuning them efficiently under black-box or memory-constrained settings remains challenging.
Approach: They propose a Riemannian zeroth-order optimization framework that constrains updates to the tangent space of the LoRA manifold.
Outcome: The proposed framework achieves more stable convergence, tighter variance bounds, and superior performance compared to existing ZO methods.
GAN Driven Semi-distant Supervision for Relation Extraction (N19-1)

Copied to clipboard

Challenge: Existing methods for relation extraction are limited to costly hand-labeled training sets and hard to be extended to large-scale relations.
Approach: They propose a semi-distant supervision approach for relation extraction by constructing a small accurate dataset and properly leveraging numerous instances without relation labels.
Outcome: The proposed approach achieves significant improvements over baselines on real-world datasets.
Active Testing: An Unbiased Evaluation Method for Distantly Supervised Relation Extraction (2020.findings-emnlp)

Copied to clipboard

Challenge: Existing methods for distantly supervised relation extraction suffer from low quality of test set, which leads to considerable biased performance evaluation.
Approach: They propose a method to evaluate distantly supervised relation extraction using noisy test sets and manual annotations.
Outcome: Experiments on a widely used benchmark show that the proposed method can yield approximately unbiased evaluations for distantly supervised relation extractors.
Understanding and Mitigating Bias Inheritance in LLM-based Data Augmentation on Downstream Tasks (2026.acl-long)

Copied to clipboard

Challenge: Generating synthetic datasets via large language models (LLMs) has emerged as promising approach to improve LLM performance.
Approach: They propose three mitigation strategies to mitigate bias inheritance in LLMs by analyzing real and LLM-augmented data.
Outcome: The proposed methods can work differently on different tasks and biases.
s1: Simple test-time scaling (2025.emnlp-main)

Copied to clipboard

Challenge: OpenAI’s o1 model showed this capability but did not publicly share its methodology, leading to many replication efforts.
Approach: They curate a small dataset s1K with 1,000 reasoning questions based on three criteria we validate through ablations: difficulty, diversity, and quality.
Outcome: The proposed model exceeds o1-preview on competition math questions by up to 27% (MATH and AIME24).
FoodieQA: A Multimodal Dataset for Fine-Grained Understanding of Chinese Food Culture (2024.emnlp-main)

Copied to clipboard

Challenge: FoodieQA is a manually curated, fine-grained image-text dataset capturing the intricate features of food cultures across various regions in China.
Approach: They evaluate vision–language Models and large language models on unseen food images and corresponding questions.
Outcome: The proposed dataset evaluates vision–language Models and large language models on unseen food images and corresponding questions.
CreditDecoding: Accelerating Parallel Decoding in Diffusion Large Language Models with Trace Credit (2026.acl-long)

Copied to clipboard

Challenge: Diffusion large language models generate text through iterative denoising with bidirectional attention, enabling richer contextual dependencies.
Approach: They propose a training-free parallel decoding method that fuses Trace Credit with current logits to boost the confidence of correct but underconfident tokens.
Outcome: The proposed method achieves 5.48 times speedup with +0.48 accuracy on LLaDA-8B and is orthogonal to mainstream inference optimizations.
Summarize-Exemplify-Reflect: Data-driven Insight Distillation Empowers LLMs for Few-shot Tabular Classification (2025.findings-emnlp)

Copied to clipboard

Challenge: Recent studies show the promise of large language models for few-shot tabular classification but highlight challenges due to the variability in structured data.
Approach: They propose a framework that distills data into actionable insights to enable robust and effective classification by large language models.
Outcome: The proposed framework integrates rule summarization, strategic exemplification, and insight reflection through deep collaboration between LLMs and data modeling techniques.
Teaching LLMs to Abstain across Languages via Multilingual Feedback (2024.emnlp-main)

Copied to clipboard

Challenge: Existing studies on LLM abstention focus on English, but they show that it can reduce the accuracy of the model by 20.5% .
Approach: They propose to teach LLMs to abstain in the face of knowledge gaps by generating multiple feedback items in related languages.
Outcome: Extensive experiments show that the proposed approach outperforms baselines and achieves 9.2% improvement for low-resource languages.

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