Papers by Xin Su

29 papers
CLEME2.0: Towards Interpretable Evaluation by Disentangling Edits for Grammatical Error Correction (2025.acl-long)

Copied to clipboard

Challenge: Existing studies have focused on the interpretability of Grammatical Error Correction (GEC) evaluation metrics, but the interpretabilty of these metrics has been neglected.
Approach: They propose a reference-based metric that describes four aspects of GEC systems: hit-correction, wrong-corrections, under-correcties, and over-corrects.
Outcome: The proposed metric reveals critical qualities and locates drawbacks of GEC systems.
Steering Away from Refusal: A Black-box Jailbreak Method Based on First-Token Distribution (2026.findings-acl)

Copied to clipboard

Challenge: Existing methods to analyze black-box jailbreaks lack direct optimization signals to refine adversarial prompts.
Approach: They propose a distribution-jailbreak attack method that selects effective jailbreak templates and iteratively optimizes adversarial suffixes by maximizing the KL divergence from the standard refusal distribution.
Outcome: The proposed method achieves state-of-the-art Attack Success Rate (ASR) on all tested open-source models and delivers over 94% ASR on GPT-4.1.
Fusing Temporal Graphs into Transformers for Time-Sensitive Question Answering (2023.findings-emnlp)

Copied to clipboard

Challenge: Existing methods for extracting temporal information from text are not suitable for time-sensitive questions.
Approach: They propose to use existing temporal information extraction systems to construct temporal graphs of events, times, and temporal relations in questions and documents.
Outcome: The proposed method outperforms graph convolution-based approaches on SituatedQA and TimeQA.
Bridging Subword Gaps in Pretrain-Finetune Paradigm for Natural Language Generation (2021.acl-long)

Copied to clipboard

Challenge: Existing methods to pretrain language models are limited by one-size-fits-all vocabulary . embeddings of mismatch tokens can be efficiently initialized in downstream tasks .
Approach: They propose to extend pretrain-finetune pipeline with an embedding transfer step . plug-and-play embeddable generator is introduced to generate any input token .
Outcome: The proposed approach allows for more efficient and better performed NLG models.
Unlocking Parameter-Efficient Fine-Tuning for Low-Resource Language Translation (2024.findings-naacl)

Copied to clipboard

Challenge: Parameter-efficient fine-tuning (PEFT) methods are important in low-resource language (LRL) Neural Machine Translation (NMT) but their practical effectiveness varies significantly across different languages.
Approach: They evaluated the performance of 8 parameters-efficient fine-tuning methods with 15 architectures using the SacreBLEU score.
Outcome: The Houlsby+Inversion adapter outperforms the baseline architectures in both in-domain and out-domain tests and the Houlson+Inverter achieves the best performance overall.
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.
StoryMI: Steerable Multi-Agent Therapeutic Dialogue Generation (2026.findings-acl)

Copied to clipboard

Challenge: Motivational interviewing (MI) is a directive, client-centered counseling approach for eliciting clients' motivation for behavioral change.
Approach: They propose a multi-LLM agent framework for controllable MI dialogue generation . therapist and client agents generate MI-coded utterances guided by MI codes .
Outcome: The proposed framework can generate fluent dialogues with minimal intervention time and a high level of evaluation.
SELECting over Tokens: Curating Pre-training Data at Scale via Token Classification (2026.acl-long)

Copied to clipboard

Challenge: Existing pipelines rely on expert-crafted heuristic rules, which lack content-aware, fine-grained noise detection.
Approach: They propose a framework that reframes data refinement as a highly efficient token classification task.
Outcome: The proposed framework outperforms existing pipelines on benchmarks and is 2.5x faster at inference.
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.
Compete to Complete: Co-opetition Adversarial Learning for Retrieval-Augmented Generation (2026.acl-long)

Copied to clipboard

Challenge: Existing approaches to reduce hallucination in large language models lack a robust mechanism for generating a generative model.
Approach: They propose a framework that formulates retriever–generator training in RAG as a minimax game.
Outcome: The proposed framework improves retrieval-augmented generation performance on seven benchmark datasets.
Read before Generate! Faithful Long Form Question Answering with Machine Reading (2022.findings-acl)

Copied to clipboard

Challenge: Long-form question answering (LFQA) generates a paragraph-length answer for a given question.
Approach: They propose a framework that jointly models answer generation and machine reading.
Outcome: The proposed model generates a more factually accurate answer from millions of documents retrieved from a large dataset.
UCS: Estimating Unseen Coverage for Improved In-Context Learning (2026.findings-acl)

Copied to clipboard

Challenge: Existing selection methods prioritize heuristic notions of relevance or diversity and provide limited insight into the coverage of a demonstration set.
Approach: They propose a training-free, subset-level coverage prior that is unrevealed by a model-consistent embedding and a Smoothed Good-Turing estimator to estimate the number of unrevelled clusters within a candidate subset.
Outcome: Experiments on multiple intent-classification and reasoning benchmarks show that augmenting strong baselines with UCS improves ICL accuracy by 2-6% under the same selection budget.
LLM-OREF: An Open Relation Extraction Framework Based on Large Language Models (2025.emnlp-main)

Copied to clipboard

Challenge: Existing studies focus on building models that can only handle predefined relations . however, their reliance on human annotation limits their practicality .
Approach: They propose an open relation extraction framework that can generalize to new relations not encountered during training.
Outcome: The proposed framework can generalize to new relations not encountered during training.
BAPO: Boundary-Aware Policy Optimization for Reliable Agentic Search (2026.findings-acl)

Copied to clipboard

Challenge: Existing RL-based agentic search models fail to recognize reasoning boundaries and rarely admit "I DON'T KNOW" lack of reliability leads to plausible but unreliable answers, introducing significant risks .
Approach: They propose a framework to cultivate reliable boundary awareness without compromising accuracy.
Outcome: Experiments show that the proposed framework improves the reliability of agentic search models.
Too Consistent to Detect: A Study of Self-Consistent Errors in LLMs (2025.emnlp-main)

Copied to clipboard

Challenge: Existing detection methods fail to account for **self-consistent error** . study identifies self-consistency errors and evaluates them .
Approach: They propose a method that fuses hidden state evidence from an external verifier LLM to detect self-consistent errors.
Outcome: The proposed method significantly enhances performance on self-consistent errors across three LLM families.
Multimodal Language Models See Better When They Look Shallower (2025.emnlp-main)

Copied to clipboard

Challenge: Existing studies show that multimodal large language models extract visual features from the final layers of a pretrained Vision Transformer.
Approach: They propose a feature fusion method that strategically incorporates shallower layers . they propose MLLMs that extract visual features from the final layers of a pretrained Vision Transformer .
Outcome: The proposed method outperforms deep layers on fine-grained visual tasks . it is the first comprehensive study of visual layer selection for MLLMs .
Getting the Most out of Simile Recognition (2022.findings-emnlp)

Copied to clipboard

Challenge: Recent work ignores features other than surface strings and suffers from data hunger issue.
Approach: They propose to use simile sentence classification and simile component extraction to find simile components.
Outcome: The proposed model outperforms current state-of-the-art systems and baselines.
Semi-Structured Chain-of-Thought: Integrating Multiple Sources of Knowledge for Improved Language Model Reasoning (2024.naacl-long)

Copied to clipboard

Challenge: Existing prompting methods rely on only one or two of these sources, or require repeatedly invoking large language models to generate similar or identical content.
Approach: They propose a semi-structured prompting approach that integrates parametric memory with unstructured knowledge from text documents and structured knowledge from knowledge graphs.
Outcome: The proposed prompting method surpasses existing prompting methods even exceeding those that require fine-tuning on open-domain multi-hop question answering datasets.
Vision-Language Introspection: Mitigating Overconfident Hallucinations in MLLMs via Interpretable Bi-Causal Steering (2026.acl-long)

Copied to clipboard

Challenge: Existing approaches to overcome object hallucination are limited . Existing mitigations include costly retraining and a training-free inference framework .
Approach: They propose a training-free inference framework that simulates a metacognitive self-correction process.
Outcome: The proposed framework reduces object hallucination rates by 12.67% on MMHal-Bench and improves accuracy by 5.8% on POPE.
TARN-VIST: Topic Aware Reinforcement Network for Visual Storytelling (2024.lrec-main)

Copied to clipboard

Challenge: Existing methods for visual storytelling ignore latent topic information.
Approach: They propose a topic-aware reinforcement network for VIsual StoryTelling that takes topic information into account to generate a coherent story.
Outcome: The proposed method outperforms most of the competing models across multiple evaluation metrics.
XL-NBT: A Cross-lingual Neural Belief Tracking Framework (D18-1)

Copied to clipboard

Challenge: a multi-lingual approach to training dialog systems is expensive and tedious, but it can be useful for cross-lingual support.
Approach: They propose to annotate data for multiple languages and train a multi-lingual dialog system for each language.
Outcome: The proposed framework bypasses the expensive human annotation and achieves promising results.
A Comparison of Strategies for Source-Free Domain Adaptation (2022.acl-long)

Copied to clipboard

Challenge: Existing research on domain adaptation without access to training data is limited due to privacy concerns.
Approach: They compare active learning, self-training, and data augmentation strategies for source-free domain adaptation with a shared task.
Outcome: The proposed algorithms yield consistent gains across all SemEval 2021 Task 10 tasks and domains, but they are unreliable for source-free domain adaptation.
Transformer-Based Temporal Information Extraction and Application: A Review (2025.emnlp-main)

Copied to clipboard

Challenge: Temporal information extraction (IE) aims to extract structured temporal information from unstructured text, thereby uncovering the implicit timelines within.
Approach: They summarize and analyze the work using Transformers to highlight potential future directions.
Outcome: The proposed method is applied across healthcare, newswire, and intelligence analysis domains.
DAST: Context-Aware Compression in LLMs via Dynamic Allocation of Soft Tokens (2025.findings-acl)

Copied to clipboard

Challenge: Existing semantic vector-based compression methods do not account for the intrinsic information density variations between context chunks, instead allocating soft tokens uniformly across context chunk.
Approach: They propose a method that leverages the LLM's intrinsic understanding of contextual relevance to guide compression.
Outcome: The proposed method surpasses state-of-the-art methods on long context tasks.
Personalized Question Answering with User Profile Generation and Compression (2025.findings-emnlp)

Copied to clipboard

Challenge: Large language models are prone to providing “midguy” answers regardless of users’ knowledge background, thereby failing to meet each user’s personalized needs.
Approach: They propose to generate personalized answers with LLMs based on users’ past question-answering records.
Outcome: The proposed method generates personalized answers based on user's past question-answering records.
PretrainRL: Alleviating Factuality Hallucination of Large Language Models at the Beginning (2026.findings-acl)

Copied to clipboard

Challenge: Large language models suffer from factual hallucinations where they generate verifiable falsehoods.
Approach: They propose a framework that integrates reinforcement learning into the pretraining phase to consolidate factual knowledge.
Outcome: The proposed framework significantly alleviates factual hallucinations and outperforms state-of-the-art methods.
Optimizing NLU Reranking Using Entity Resolution Signals in Multi-domain Dialog Systems (2021.naacl-industry)

Copied to clipboard

Challenge: In dialog systems, the Natural Language Understanding component makes the interpretation decision before the mentioned entities are resolved.
Approach: They propose to leverage Entity Resolution (ER) features in NLU reranking to learn model weights . they propose a score distribution matching method to ensure the models are calibrated .
Outcome: The proposed approach outperforms the baseline model on multiple domain evaluations.
BaseCal: Unsupervised Confidence Calibration via Base Model Signals (2026.acl-long)

Copied to clipboard

Challenge: Post-trained LLMs typically compromise reliability with severe overconfidence, resulting in inaccurate responses.
Approach: They propose a solution that feeds PoLLMs into the base LLM to get confidence.
Outcome: The proposed solution reduces expected calibration error (ECE) by 42.90% compared to the best unsupervised baselines.
IBADR: an Iterative Bias-Aware Dataset Refinement Framework for Debiasing NLU models (2023.emnlp-main)

Copied to clipboard

Challenge: Using manual data analysis, dataset refinement approaches are often unable to cover all the potential biased features.
Approach: They propose an iterative bias-aware dataset refinement framework which debiases NLU models without predefining biased features.
Outcome: The proposed framework outperforms existing methods and is compatible with model-centric methods.

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