Papers by Xiaoming Liu

28 papers
RepoDebug: Repository-Level Multi-Task and Multi-Language Debugging Evaluation of Large Language Models (2025.findings-emnlp)

Copied to clipboard

Challenge: Large Language Models (LLMs) have exhibited significant proficiency in code debugging, especially in automatic program repair.
Approach: They propose a repository-level code debugging dataset with 22 subtypes of errors that supports 8 commonly used programming languages and 3 debug tasks.
Outcome: The proposed dataset supports 8 commonly used programming languages and 3 debugging tasks.
CoCo: Coherence-Enhanced Machine-Generated Text Detection Under Low Resource With Contrastive Learning (2023.emnlp-main)

Copied to clipboard

Challenge: Recent proposed methods fail to consider the linguistic structure of texts and lack the ability to handle the low-resource problem.
Approach: They propose a coherence-based contrastive learning model named CoCo to detect MGTs under low-resource scenario.
Outcome: The proposed model outperforms state-of-the-art methods on two datasets and two self-constructed datasets.
PDC & DM-SFT: A Road for LLM SQL Bug-Fix Enhancing (2025.coling-industry)

Copied to clipboard

Challenge: Existing code large language models focus on generating correct code, but struggle with bug repair.
Approach: They propose a set of methods to enhance LLM’s SQL bug-fixing abilities by combining a data set construction and a supervised bug-fixed learning approach.
Outcome: The proposed methods exceed current best performing model which size is much larger.
Stumbling Blocks: Stress Testing the Robustness of Machine-Generated Text Detectors Under Attacks (2024.acl-long)

Copied to clipboard

Challenge: Existing studies on this topic focus on the robustness of specific detectors or particular attack methods.
Approach: They stress test the detectors’ robustness to malicious attacks under realistic scenarios using LLMs and metric-based detectors.
Outcome: The proposed methods are based on a set of LLM-based models and their performance is compared under different budget levels.
Matching Distributions between Model and Data: Cross-domain Knowledge Distillation for Unsupervised Domain Adaptation (2021.acl-long)

Copied to clipboard

Challenge: Existing methods require to learn to adapt the target model by exploiting the source data and sharing the network architecture across domains.
Approach: They propose a framework that allows to transfer the knowledge of source domain to the unlabeled target domain without using source data.
Outcome: The proposed framework matches distributions between a trained source model and a set of target data and achieves superior performance on cross-domain text classification.
AgentCPM-GUI: Building Mobile-Use Agents with Reinforcement Fine-Tuning (2025.emnlp-demos)

Copied to clipboard

Challenge: Large language model agents have enabled GUI-based automation, but their deployment is limited by noisy data, poor generalization, and lack of support for non-English GUIs.
Approach: They propose an 8B-parameter GUI agent built for robust and efficient on-device GUI interaction.
Outcome: The proposed GUI agent achieves promising performance on five public benchmarks and proposed Chinese benchmark CAGUI.
Towards Database-Free Text-to-SQL Evaluation: A Graph-Based Metric for Functional Correctness (2025.coling-main)

Copied to clipboard

Challenge: Existing metrics for evaluating functional correctness of SQL queries are prone to false positives due to inadequately prepared test databases.
Approach: They propose a graph-based metric that uses a relational operator tree to extract rich semantic information from the logical execution plan of SQL queries and embed it into a diagram.
Outcome: The proposed method eliminates the need for extensive test database preparation and performs graph matching on unseen SQL queries.
Exploiting Contextual Information via Dynamic Memory Network for Event Detection (D18-1)

Copied to clipboard

Challenge: Existing methods for event detection only process context once . a multi-hop mechanism to capture contextual information improves performance .
Approach: They propose to use dynamic memory network to capture contextual information . they propose to model event triggers by identifying word or phrase which most represents it .
Outcome: The proposed model achieves best F1 score compared to the state-of-the-art models.
Iron Sharpens Iron: Defending Against Attacks in Machine-Generated Text Detection with Adversarial Training (2025.acl-long)

Copied to clipboard

Challenge: Existing MGT detectors are vulnerable to simple perturbations and adversarial attacks.
Approach: They propose an adversarial framework for training a robust machine-generated text detector called GREedy Adversary PromoTed DefendER.
Outcome: The proposed framework reduces the Attack Success Rate (ASR) by 0.67% compared with SOTA defense methods.
KwaiChat: A Large-Scale Video-Driven Multilingual Mixed-Type Dialogue Corpus (2025.findings-naacl)

Copied to clipboard

Challenge: Currently, video-based dialogue systems rely on a single dialogue type, hindering their versatility in practical applications.
Approach: They propose to generate video-driven multilingual mixed-type dialogues using KwaiChat . they propose to create a video-based multilingual mix of 4 dialogue types, 30 domains, 4 languages, 13 topics .
Outcome: The proposed model performs best on KwaiChat but is not perfect in this situation.
Improving Pacing in Long-Form Story Planning (2023.findings-emnlp)

Copied to clipboard

Challenge: Existing systems for writing long-form stories suffer from unnatural pacing, whether glossing over important events or over-elaborating on insignificant details.
Approach: They propose a system that uses a concreteness evaluator to judge which of two events is more concrete.
Outcome: The proposed system improves pacing when automatically generating story outlines.
DataSage: Multi-agent Collaboration for Insight Discovery with External Knowledge Retrieval, Multi-role Debating, and Multi-path Reasoning (2026.findings-acl)

Copied to clipboard

Challenge: Existing data insight agents fail to deliver satisfactory results due to insufficient utilization of domain knowledge, shallow analytical depth, and error-prone code generation.
Approach: They propose a novel multi-agent framework that incorporates external knowledge retrieval to enrich the analytical context, a multi-role debating mechanism to simulate diverse analytical perspectives and deepen analytical depth, and multi-path reasoning to improve the accuracy of the generated code and insights.
Outcome: Extensive experiments on InsightBench show that DataSage outperforms existing data insight agents across all difficulty levels, improving by 7.5% and 13.9% respectively in insight-level and summary-level metrics.
Confidence Should Be Calibrated More Than One Turn Deep (2026.acl-long)

Copied to clipboard

Challenge: Existing work on confidence estimation and calibration focuses on single-turn settings . existing work on multi-turn calibration ignores the risks and potential of multi-turned conversations .
Approach: They propose a multi-turn calibration task that reframes calibration from a static property into a dynamic challenge central to reliable multi- turn conversations.
Outcome: The proposed model minimizes ECE@T and leverages ConfChat to improve confidence . the proposed model preserves and even enhances model performance in multi-turn interactions.
Stealthy Jailbreak Attacks on Large Language Models via Benign Data Mirroring (2025.naacl-long)

Copied to clipboard

Challenge: Existing black-box jailbreak methods often rely on model feedback . existing methods may be intercepted by content moderators during the search process .
Approach: They propose a method that guides malicious prompt construction by local training a mirror model of the target black-box model through benign data distillation.
Outcome: The proposed method achieves a 92% attack success rate and 80% stealth rate on a subset of AdvBench.
Graph-Reward-SQL: Execution-Free Reinforcement Learning for Text-to-SQL via Graph Matching and Stepwise Reward (2025.findings-emnlp)

Copied to clipboard

Challenge: Existing methods to enhance performance of large language models (LLMs) on Text-to-SQL tasks rely on execution-based or LLM-based reward models.
Approach: They propose a reward model framework for RL-based Text-to-SQL that employs the GMNScore outcome reward model.
Outcome: The proposed reward model outperforms existing reward models on standard benchmarks including Spider and BIRD.
Beyond Static Persona Consistency: Dynamic Persona Coherence in LLM Role-Playing (2026.acl-long)

Copied to clipboard

Challenge: Existing LLMs conflate identity consistency with emotional rigidity . Existing models exhibit either robotic repetition or persona drift .
Approach: They propose a framework that decouples Identity-Layer Stability from Adaptive-Layer Appropriateness to achieve persona coherence repair.
Outcome: Experiments on GPT-4o, Claude-3.5-Sonnet, and DeepSeek-V3.2 show consistent improvements (+16–84% gains)
Direct Large Language Model Alignment Through Self-Rewarding Contrastive Prompt Distillation (2024.acl-long)

Copied to clipboard

Challenge: Existing methods to evaluate preference data without human annotations are difficult . et al., 2022b) is effective for aligning large language models with human expectations .
Approach: They propose a method to evaluate the response preference using output probabilities under contrastive prompts.
Outcome: The proposed method could surpass the RLHF method without human-annotated preference data.
Can Reasoning Path still be Effective as Input? Bridging Post-Reasoning to Chain-of-Thought Compression (2026.acl-long)

Copied to clipboard

Challenge: Existing work on reducing CoT generation in reasoning impairs the necessary information for deriving the correct answer.
Approach: They propose a reasoning paradigm that takes CoT as a part of context to simplify the reasoning task for Large Language Models (LLMs).
Outcome: The proposed framework reduces the generation length of LLMs, but its effectiveness hinges on the efficiency and reliability of the contextual CoT generation.
Medical Dialogue System: A Survey of Categories, Methods, Evaluation and Challenges (2024.findings-acl)

Copied to clipboard

Challenge: Existing medical dialogue systems have significant potential to simplify diagnostic procedure and reduce the cost of collecting information from patients.
Approach: They analyze 325 papers from well-known computer science, natural language processing conferences and journals to find out the major challenges of medical dialog systems.
Outcome: The proposed systems have been surveyed in the medical community but have not been evaluated from a technical perspective.
MidMed: Towards Mixed-Type Dialogues for Medical Consultation (2023.acl-long)

Copied to clipboard

Challenge: Current medical dialogue systems assume that patients have explicit goals but are often unavailable in real-world situations due to the lack of medical knowledge.
Approach: They propose a human-to-human mixed-type medical consultation dialogue corpus . they build benchmarking baselines on MidMed and propose an instruction-guiding framework . Experimental results show the effectiveness of InsMed .
Outcome: The proposed system can help patients clarify their goals in real-world situations . it covers four departments with 8,309 dialogues and provides benchmarking baselines .
UniDataBench: Evaluating Data Analytics Agents Across Structured and Unstructured Data (2026.acl-long)

Copied to clipboard

Challenge: Existing benchmarks do not assess agents’ capabilities across data types . Existing tools only evaluate agents' ability to extract reasonable insights across data formats.
Approach: They propose a multi-source benchmark to evaluate the performance of data analytics agents in handling diverse data sources.
Outcome: The proposed agent performs end-to-end analysis over diverse data sources by automatically discovering cross-source linkages, decomposing goals, and generating robust, self-correcting code to extract actionable insights.
Does DetectGPT Fully Utilize Perturbation? Bridging Selective Perturbation to Fine-tuned Contrastive Learning Detector would be Better (2024.acl-long)

Copied to clipboard

Challenge: Existing methods to detect MGT from human-written texts are inadequate . existing methods are fine-tuned and zero-shot metric-based, but they can be more accurate.
Approach: They propose a novel fine-tuned detector that can detect MGT from human-written texts by contrastive learning on selective perturbation.
Outcome: The proposed method outperforms the state-of-the-art by 1.20% on four public datasets.
Mis-prompt: Benchmarking Large Language Models for Proactive Error Handling (2025.acl-long)

Copied to clipboard

Challenge: Current error-handling works are performed in a passive manner, with explicit error- handling instructions.
Approach: They propose a new benchmark to analyze LLMs' performance on a mis-prompt benchmark and a dataset to promote further research.
Outcome: The proposed benchmark shows that current LLMs show poor performance on proactive error handling, and that SFT improves on error handling instances.
LIFBench: Evaluating the Instruction Following Performance and Stability of Large Language Models in Long-Context Scenarios (2025.acl-long)

Copied to clipboard

Challenge: Existing benchmarks rarely focus on instruction-following in long-context scenarios or stability on different inputs.
Approach: They propose a scalable dataset to evaluate LLMs’ instruction-following capabilities and stability across long contexts.
Outcome: The proposed method evaluates LLMs’ instruction-following capabilities and stability across long contexts.
FinEval: A Chinese Financial Domain Knowledge Evaluation Benchmark for Large Language Models (2025.naacl-long)

Copied to clipboard

Challenge: Large language models have demonstrated outstanding performance in various natural language processing tasks, but their security capabilities in the financial domain have not been explored.
Approach: They propose to use a benchmark to evaluate large language models' financial domain knowledge and practical abilities.
Outcome: The proposed benchmark evaluates large language models' financial domain knowledge and practical abilities.
Flow2Code: Evaluating Large Language Models for Flowchart-based Code Generation Capability (2025.findings-acl)

Copied to clipboard

Challenge: Existing code generation benchmarks neglect flowchart-based code generation . existing benchmarks lack flowcharting-based evaluation, limiting the potential of large language models and minimizing human error.
Approach: They propose to use flowcharts to evaluate existing LLMs' code generation capabilities.
Outcome: The proposed benchmarks show that the supervised fine-tuning technique contributes greatly to the models’ performance.
MDS: A Fine-Grained Dataset for Multi-Modal Dialogue Summarization (2024.lrec-main)

Copied to clipboard

Challenge: Summarizing the dialogue into a short message has drawn much attention due to the explosion of various dialogue scenes.
Approach: They develop a multi-modal dialogue summarization dataset to enhance the variety of data available for this research area.
Outcome: The proposed dataset provides a demanding testbed for multi-modal dialogue summarization.
StablePT : Towards Stable Prompting for Few-shot Learning via Input Separation (2024.findings-emnlp)

Copied to clipboard

Challenge: Existing studies on prompt tuning have shown that language models can be effective few-shot learners with prompting.
Approach: They propose to treat the hard prompt and soft prompt as separate inputs to mitigate noise brought by prompt initialization.
Outcome: Experimental results show that the proposed method outperforms state-of-the-art methods by 6.97% in accuracy and reduces the standard deviation by 1.92 on average.

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