Papers by Lei Xue

21 papers
Parameter Importance is Not Static: Evolving Parameter Isolation for Supervised Fine-Tuning (2026.acl-long)

Copied to clipboard

Challenge: Recent approaches to fine-tuning of large language models suffer from task interference and catastrophic forgetting.
Approach: They propose a fine-tuning framework that adapts isolation decisions based on online estimates of parameter importance.
Outcome: The proposed framework reduces interference and forgetting while releasing outdated parameters to recover plasticity.
Why Supervised Fine-Tuning Fails to Learn: A Systematic Study of Incomplete Learning in Large Language Models (2026.acl-long)

Copied to clipboard

Challenge: Incomplete learning is widespread and heterogeneous in large language models . authors identify five recurrent sources of incomplete learning: missing prerequisite knowledge, conflicts between SFT supervision and pre-training knowledge, internal inconsistencies within SFT data, left-side forgetting during sequential fine-tuning, and insufficient optimization for rare or complex patterns.
Approach: They propose a diagnostic-first framework that maps incomplete learning to causes . they identify five recurrent sources of incomplete learning: missing prerequisite knowledge, conflicts between supervision and pre-training knowledge, internal inconsistencies, left-side forgetting during sequential fine-tuning, and insufficient optimization for rare or complex patterns.
Outcome: The proposed framework maps incomplete learning to causes using observable training and inference signals.
Unifying Discrete and Continuous Representations for Unsupervised Paraphrase Generation (2023.emnlp-main)

Copied to clipboard

Challenge: Existing unsupervised paraphrase generation methods require large-scale, manually annotated paraphrase datasets, which are labor-intensive to build.
Approach: They propose a self-supervised pseudo-data construction method that generates diverse pseudo-paraphrases in distinct surface structures for a given sentence.
Outcome: The proposed method generates diverse pseudo-paraphrases in distinct surface structures for a given sentence.
Attention Consistency for LLMs Explanation (2025.findings-emnlp)

Copied to clipboard

Challenge: Existing interpretability methods face limitations such as low resolution and high computational cost.
Approach: They propose a multi-layer attention consistency score to estimate the importance of input tokens in large language models.
Outcome: The proposed heuristic achieves a favorable trade-off between interpretability quality and computational efficiency .
Reason Only When Needed: Efficient Generative Reward Modeling via Model-Internal Uncertainty (2026.findings-acl)

Copied to clipboard

Challenge: Existing approaches to generating reward models rely on voting-based mechanisms to evaluate CoT outputs.
Approach: They propose an efficient generative reward modeling framework grounded in model-internal uncertainty.
Outcome: The proposed framework reduces inference cost while improving answer accuracy.
GCPG: A General Framework for Controllable Paraphrase Generation (2022.findings-acl)

Copied to clipboard

Challenge: Existing studies highlight a special condition under two indispensable aspects of controllable paraphrase generation (CPG) individually, lacking a unified circumstance to explore and analyze their effectiveness.
Approach: They propose a general controllable paraphrase generation framework that integrates lexical and syntactical conditions into a text sequence and uniformly processes them in an encoder-decoder paradigm.
Outcome: The proposed framework can combine lexical and syntactical conditions and improve paraphrase generation.
Language Generation via Combinatorial Constraint Satisfaction: A Tree Search Enhanced Monte-Carlo Approach (2020.findings-emnlp)

Copied to clipboard

Challenge: Generating natural language under complex constraints is a principled formulation towards controllable text generation.
Approach: They propose a method to specify combinatorial constraints for sentence generation . they use a tree search algorithm embedded into the proposal process of the Markov Chain Monte Carlo .
Outcome: The proposed method achieves consistent and significant improvement on multiple language generation tasks.
Tailor: A Soft-Prompt-Based Approach to Attribute-Based Controlled Text Generation (2023.acl-long)

Copied to clipboard

Challenge: Existing work focuses on generating sentences satisfying pre-specified attributes such as topic and sentiment, yet suffers from increases in storage and inference time.
Approach: They propose a method that uses a pre-trained continuous vector to generate a fixed pre-trainable language model to satisfy a specified attribute.
Outcome: The proposed model can achieve improvements on eleven attribute-specific generation tasks with 0.08% extra training parameters.
ReSo: A Reward-driven Self-organizing LLM-based Multi-Agent System for Reasoning Tasks (2025.emnlp-main)

Copied to clipboard

Challenge: Multi-agent systems (MAS) are limited by poor flexibility and scalability, with underdeveloped optimization strategies.
Approach: They propose a task graph generation and a reward-driven two-stage agent selection process to integrate multi-agent systems to improve their reasoning capabilities.
Outcome: The proposed model outperforms existing methods on Math-MAS and SciBench-MAS SciBech, while other methods completely fail.
LLaSE-G1: Incentivizing Generalization Capability for LLaMA-based Speech Enhancement (2025.acl-long)

Copied to clipboard

Challenge: Recent advances in language models have demonstrated strong capabilities in semantic understanding and contextual modeling.
Approach: They propose a LLaMA-based language model that incentivizes generalization capabilities for speech enhancement.
Outcome: The proposed language model outperforms prior task-specific discriminative and generative models in acoustic enhancement tasks.
Dynamic Voting for Efficient Reasoning in Large Language Models (2023.findings-emnlp)

Copied to clipboard

Challenge: Multi-path voting methods generate multiple reasoning paths for each problem, causing factual errors and illusion generation.
Approach: They propose a multi-path voting technique that effectively reduces the number of reasoning paths during multi-path voting while preserving accuracies.
Outcome: The proposed method outperforms Self-consistency using 24.7% of the number of paths on the LetterConcat task.
CKDST: Comprehensively and Effectively Distill Knowledge from Machine Translation to End-to-End Speech Translation (2023.findings-acl)

Copied to clipboard

Challenge: End-to-end speech-totext translation (ST) data are limited due to the limited resources.
Approach: They propose a knowledge distillation framework for speech translation that integrates knowledge from machine translation and decouples knowledge from non-target class knowledge.
Outcome: The proposed framework outperforms state-of-the-art models on a benchmark dataset.
WenetSpeech-Wu: Datasets, Benchmarks, and Models for a Unified Chinese Wu Dialect Speech Processing Ecosystem (2026.findings-acl)

Copied to clipboard

Challenge: despite its linguistic significance, the Wu dialect of Chinese has long been hindered by the lack of large-scale speech data, standardized evaluation benchmarks, and publicly available models.
Approach: They propose to use WenetSpeech-Wu as a large-scale, multi-dimensionally annotated open-source speech corpus for the Wu dialect of Chinese.
Outcome: The proposed dataset includes 8,000 hours of speech data and strong open-source models . the proposed dataset is competitive and empirically validated .
AtTGen: Attribute Tree Generation for Real-World Attribute Joint Extraction (2023.acl-long)

Copied to clipboard

Challenge: Attribute extraction aims to identify attribute names and the corresponding attribute values from descriptive texts.
Approach: They propose a unified formulation for real-world attribute extraction application, where closed-world, open-world and semi-open attribute extraction tasks are modeled uniformly.
Outcome: The proposed model outperforms existing methods on three datasets and outperformed existing methods by a large margin.
BayesKD: Bayesian Knowledge Distillation for Compact LLMs in Constrained Fine-tuning Scenarios (2025.findings-acl)

Copied to clipboard

Challenge: Large language models (LLMs) have revolutionized various domains with their remarkable capabilities, but their massive parameter sizes pose significant challenges for fine-tuning and inference.
Approach: They propose a Bayesian Knowledge Distillation framework for compact Large Language Models in resource-constrained fine-tuning scenarios that employs Logits Dual-Scaling, Knowledge Alignment Module, and Bayes Distillations Optimization.
Outcome: The proposed framework outperforms baseline methods on various state-of-the-art LLMs, including LLaMA, Qwen2, Bloom, and Vicuna.
How Good Are LLMs at Out-of-Distribution Detection? (2024.lrec-main)

Copied to clipboard

Challenge: Out-of-distribution (OOD) detection is crucial for ensuring AI safety . large language models (LLMs) are becoming more prevalent due to their scale, pre-training objectives, and paradigms used for inference.
Approach: They propose to use large language models to investigate out-of-distribution (OOD) detection in machine learning.
Outcome: The proposed method outperforms other OOD detectors in zero-grad and fine-tuning scenarios.
TrojFSP: Trojan Insertion in Few-shot Prompt Tuning (2024.naacl-long)

Copied to clipboard

Challenge: Prompt tuning on a few data samples presents security issues, e.g., Trojan attacks.
Approach: They propose a method to transfer established data poisoning attacks directly to few-shot prompt tuning, a technique to address the poisoned imbalance issue.
Outcome: The proposed method achieves an ASR of over 99% while maintaining negligible decreases in CDA.
Scaling Behaviors of LLM Reinforcement Learning Post-Training: An Empirical Study in Mathematical Reasoning (2026.acl-long)

Copied to clipboard

Challenge: elucidating scaling laws for large language models (LLMs) during pre-training remains unexplored.
Approach: They characterize how model scale, data, and compute interact during pre-training . they find that large models consistently demonstrate superior compute and data efficiency .
Outcome: The proposed scaling laws offer practical guidance for scaling reasoning capabilities through reinforcement learning post-training.
Self-attention-based Graph-of-Thought for Math Problem Solving (2025.findings-acl)

Copied to clipboard

Challenge: Existing methods for generating reasoning paths in a chain structure are inefficient and non-human-like.
Approach: They propose a decoding method for a chain-based LLM that constructs a thought graph simultaneously as an LLM inference and generates reasoning steps with a graph-structured self-attention mechanism.
Outcome: The proposed method improves reasoning accuracy without huge computational over-expensive LLMs and avoids performance degradation issues when the LLM is too small to comprehend complex prompts.
Unified Hallucination Detection for Multimodal Large Language Models (2024.acl-long)

Copied to clipboard

Challenge: despite significant strides in multimodal tasks, MLLMs are plagued by the critical issue of hallucination.
Approach: They propose a meta-evaluation benchmark to facilitate evaluation of advancements in hallucination detection methods.
Outcome: The proposed framework validates hallucinations robustly and provides strategic insights . MHaluBench is a meta-evaluation benchmark designed to facilitate evaluation .
Beyond Layout Embedding: Layout Attention with Gaussian Biases for Structured Document Understanding (2023.findings-emnlp)

Copied to clipboard

Challenge: Existing methods for encoding layout information rely on millions of learnable parameters . polar coordinates provide superior choice for layout modeling, study finds .
Approach: They propose to model layout attention with Gaussian biases by feeding polar coordinates into 2-D Gausssian kernels.
Outcome: The proposed model improves on three widely used benchmarks.

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