Papers by Xue Peng

23 papers
SGA-MCTS: Decoupling Planning from Execution via Training-Free Atomic Experience Retrieval (2026.findings-acl)

Copied to clipboard

Challenge: a new framework casts LLM planning as non-parametric retrieval, but high latency of inference-time search and supervised fine-tuning are limitations.
Approach: They propose a framework that casts LLM planning as non-parametric retrieval . they leverage Monte Carlo Tree Search to explore the solution space .
Outcome: Empirical results show that SGA-MCTS can match the performance of SOTA systems without task-specific fine-tuning.
RubricBench: Aligning Model-Generated Rubrics with Human Standards (2026.acl-long)

Copied to clipboard

Challenge: Existing benchmarks lack discriminative complexity and ground-truth rubric annotations required for rigorous evaluation.
Approach: They propose a curated benchmark with 1,147 pairwise comparisons to assess the reliability of rubric-based evaluation.
Outcome: The proposed benchmarks show that they support diverse domains, exhibit discriminative ability, provide high-quality annotations, and include human-authored rubrics.
A Semi-supervised Scalable Unified Framework for E-commerce Query Classification (2025.acl-industry)

Copied to clipboard

Challenge: Existing query classification methods rely on posterior click behavior to construct training samples, resulting in insufficient prior information for modeling.
Approach: They propose a semi-supervised scaleable unified framework that integrates enhanced modules to unify query classification tasks.
Outcome: The proposed framework outperforms the state-of-the-art models in offline and online A/B experiments.
Improving Preference Alignment of LLM with Inference-Free Self-Refinement (2025.findings-emnlp)

Copied to clipboard

Challenge: Large language models (LLMs) develop in-context learning capability through pretraining and instruction tuning.
Approach: Large language models (LLMs) develop in-context learning capability through pretraining and instruction tuning.
Outcome: Experiments show that incorporating IFSR into preference alignment yields performance improvement over 10%.
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.
ℛ3: Advertisement Compliance ℛectification via Group-ℛelative Experience Extractor and Curriculum ℛeinforcement (2026.acl-industry)

Copied to clipboard

Challenge: Existing methods of content moderation are infeasible due to over-editing and compromise the advertiser’s original semantic intent.
Approach: They propose a framework to harmonize compliance with original intent preservation that integrates a data-driven framework and a curriculum to enforce compliance while maximizing semantic consistency.
Outcome: The proposed framework outperforms state-of-the-art baselines on industrial datasets and on online A/B testing on industrial video.
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.
Argument-Aware Approach To Event Linking (2024.findings-acl)

Copied to clipboard

Challenge: Prior research in event linking has mainly borrowed methods from entity linking, overlooking distinct features of events.
Approach: They propose an argument-aware method to improve event linking models by augmenting input text with tagged event argument information.
Outcome: The proposed method improves in-KB and out-of-KB queries and training examples.
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.
Same Claim, Different Judgment: Benchmarking Scenario-Induced Bias in Multilingual Financial Misinformation Detection (2026.findings-acl)

Copied to clipboard

Challenge: Existing research on LLM biases has focused on direct questioning or general-purpose settings . pronounced behavioral biase despite their growing deployment in financial analysis, forecasting, and decision support.
Approach: They propose a benchmark to evaluate behavioral biases of large language models in MFMD . they use a multilingual financial misinformation dataset to integrate these with misinformation claims .
Outcome: The proposed benchmark evaluates behavioral biases of large language models across economic scenarios.
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.
Large Language Models Are Cross-Lingual Knowledge-Free Reasoners (2025.naacl-long)

Copied to clipboard

Challenge: Large language models have demonstrated impressive reasoning capabilities across multiple languages, but the relationship between capabilities in different languages is less explored.
Approach: They decompose the process of reasoning tasks into two separate components: knowledge retrieval and knowledge-free reasoning.
Outcome: The proposed model can be transferred across source-target languages despite secondary impact of resource in some specific target languages, while cross-lingual knowledge retrieval significantly hinders the transfer.
FinReporting: An Agentic Workflow for Localized Reporting of Cross-Jurisdiction Financial Disclosure (2026.acl-demo)

Copied to clipboard

Challenge: FinReporting is an agentic workflow for localized cross-jurisdiction financial reporting . existing approaches assume a single-market setting and overlook structural differences across jurisdictions .
Approach: They propose a workflow that decomposes financial reporting into auditable stages . they use Large Language Models to extract and summarize corporate disclosures .
Outcome: The proposed system decomposes reporting into auditable stages . it improves consistency and reliability under heterogeneous reporting regimes.
ImF: Embedding an Implicit Fingerprint in Your Large Language Models (2026.acl-long)

Copied to clipboard

Challenge: Training and serving large language models (LLMs) is resource-intensive, making reliable intellectual property protection and black-box ownership verification increasingly important.
Approach: They propose a method to inject a small set of secret query–response behaviors into model fingerprinting . they encode ownership information into a natural-looking target response and derive a semantically aligned query .
Outcome: The proposed fingerprints improve stealthiness and remain verifiable under model updates and deployment-time prompt interventions.
A2ATS: Retrieval-Based KV Cache Reduction via Windowed Rotary Position Embedding and Query-Aware Vector Quantization (2025.findings-acl)

Copied to clipboard

Challenge: Long context large language models (LLMs) pose significant challenges for efficient serving due to the large memory footprint and high access overhead of KV cache.
Approach: They propose a retrieval-based method to reduce the memory footprint of LLMs . they propose Windowed Rotary Position Embedding and query-aware vector quantization .
Outcome: The proposed method can achieve lower performance degradation with lower overhead compared to existing methods . it can reduce the memory footprint and access overhead of long context large language models .
SLIP: Soft Label Mechanism and Key-Extraction-Guided CoT-based Defense Against Instruction Backdoor in APIs (2026.findings-acl)

Copied to clipboard

Challenge: Existing black-box instruction backdoors can detect poisoned inputs, but fail to recover correct outputs once the backdoor is activated.
Approach: They propose a soft label mechanism and key-extraction-guided CoT-based defense against instruction backdoors in APIs (SLIP) they propose KCOT-based model to extract task-relevant keywords and phrases rather than only considering the single trigger or overall text semantics.
Outcome: The proposed model reduces the average attack success rate to 25.13% and improves clean accuracy to 87.15% and outperforms state-of-the-art black-box defenses.
Kill two birds with one stone: generalized and robust AI-generated text detection via dynamic perturbations (2025.naacl-long)

Copied to clipboard

Challenge: Existing methods focus on model generalization or focus on robustness.
Approach: They propose a model-based AIGT detection method that can be generalized and robust under two adversarial attacks.
Outcome: The proposed method outperforms state-of-the-art methods for generalization and robustness under two text adversarial attacks.
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.
metaCAT: A Metadata-based Task-oriented Chatbot Annotation Tool (2020.aacl-demo)

Copied to clipboard

Challenge: Creating high-quality annotated dialogue corpora necessitates a high level of human engagements.
Approach: They propose to develop an annotation tool specifically for developing task-oriented dialogue data that provides comprehensive metadata annotation coverage to the domain, intent, and span information.
Outcome: The tool provides comprehensive metadata annotation coverage to domain, intent, and span information.
Inhibitory Attacks on Backdoor-based Fingerprinting for Large Language Models (2026.acl-long)

Copied to clipboard

Challenge: Backdoor-based LLM fingerprinting is a promising solution for intellectual property protection . however, the vulnerability of existing LLMs for the ensemble scenario is unexplored .
Approach: They propose two new fingerprinting attack methods to assess the robustness of LLM fingerprinting by token filter attack and sentence verification attack.
Outcome: The proposed methods inhibit the fingerprint response while maintaining ensemble performance.
SSR-A: Spatial- and Semantic-Aware Instructions and Curriculum Reinforcement for Advertisement Compliant Rectification (2026.acl-industry)

Copied to clipboard

Challenge: Existing methods to fix non-compliant images suffer from over-editing, destroying original intent and perceptual similarity.
Approach: They propose a framework for the minimalist rectification of non-compliant image ads.
Outcome: The proposed framework outperforms state-of-the-art baselines in both compliance and preservation of visual and commercial consistency.
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.
Semi-supervised Domain Adaptation for Dependency Parsing (P19-1)

Copied to clipboard

Challenge: Currently, most studies on cross-domain parsing focus on unsupervised domain adaptation . however, unsupervised approaches make limited progress due to the intrinsic difficulty of both domain adaptation and parse.
Approach: They propose a semi-supervised domain adaptation problem for Chinese dependency parsing by using newly-annotated large-scale domain-aware datasets.
Outcome: The proposed method is more effective than direct corpus concatenation and multi-task learning.

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