Papers by Heuiseok Lim

Semantic Aware Linear Transfer by Recycling Pre-trained Language Models for Cross-lingual Transfer (2025.findings-acl)

Challenge:	Korean pretrained language models struggle to generate short sentences with a given condition based on compositionality and commonsense reasoning.
Approach:	They propose a Korean text-generation dataset for Korean generative commonsense reasoning and language model evaluation using a semi-automatic dataset construction approach.
Outcome:	The proposed dataset is available at http://aihub.or.kr/opendata/korea-university.

Copied to clipboard

Seungyoon Lee, Seongtae Hong, Hyeonseok Moon, Heuiseok Lim

Detecting Critical Errors Considering Cross-Cultural Factors in English-Korean Translation (2024.lrec-main)

Challenge:	Large Language Models (LLMs) are increasingly incorporating multilingual capabilities, fueling the demand to transfer them into target language-specific models.
Approach:	They propose a novel cross-lingual transfer technique that recycles embeddings from target language Pre-trained Language Models to transmit deep representational strengths to LLMs.
Outcome:	The proposed technique outperforms existing methods in cross-lingual understanding setups and achieves faster convergence and lower loss during language adaptation.

Copied to clipboard

Sugyeong Eo, Jungwoo Lim, Chanjun Park, DaHyun Jung, Seonmin Koo, Hyeonseok Moon, Jaehyung Seo, Heuiseok Lim

Machine Translation

Challenge:	Recent machine translation systems overcome language barriers for a wide range of users, yet they carry the risk of catastrophic meaning deviations.
Approach:	They introduce a culture-aware "Politeness" type for detecting critical translation errors . they also provide multiclass labels for critical error detection and critical error type classification .
Outcome:	Empirical results show that the proposed method outperforms baselines in both tasks.

Call for Rigor in Reporting Quality of Instruction Tuning Data (2025.acl-short)

Copied to clipboard

Hyeonseok Moon, Jaehyung Seo, Heuiseok Lim

LimaCost: Data Valuation for Instruction Tuning of Large Language Models (2025.findings-emnlp)

Challenge:	Instruction tuning is crucial for adapting large language models (LLMs) to user intentions.
Approach:	They propose to use hyperparameters for training models that are often selected arbitrarily without adequate justification to make arbitrary conclusions.
Outcome:	The results show that arbitrary hyperparameter decisions can make any arbitrary conclusion.

Copied to clipboard

Hyeonseok Moon, Jaehyung Seo, Seonmin Koo, Jinsung Kim, Young-kyoung Ham, Jiwon Moon, Heuiseok Lim

Where am I? Large Language Models Wandering between Semantics and Structures in Long Contexts (2024.emnlp-main)

Challenge:	Instruction tuning is an effective approach for aligning large language models with human intentions.
Approach:	They propose a data quality measure that exhibits a strong correlation with model performance.
Outcome:	The proposed measure exhibits a strong correlation with model performance.

Copied to clipboard

Seonmin Koo, Jinsung Kim, YoungJoon Jang, Chanjun Park, Heuiseok Lim

Length-aware Byte Pair Encoding for Mitigating Over-segmentation in Korean Machine Translation (2024.findings-acl)

Challenge:	Existing evaluations of the open-domain question answering task focus solely on whether the model provides the correct answer.
Approach:	They propose to examine the phenomenon of discrepancies in abilities across two distinct tasks—QA and evidence selection—when performed simultaneously.
Outcome:	The proposed framework and resources examines the ability of large language models to perform two distinct tasks simultaneously, from the perspective of task alignment.

Copied to clipboard

Jungseob Lee, Hyeonseok Moon, Seungjun Lee, Chanjun Park, Sugyeong Eo, Hyunwoong Ko, Jaehyung Seo, Seungyoon Lee, Heuiseok Lim

Machine Translation

Challenge:	Byte Pair Encoding (BPE) is an effective approach in machine translation across several languages, but it is prone to over-segmentation in Korean, an agglutinative and morphologically rich language.
Approach:	They propose a new method that incorporates long words into the Korean vocabulary by strategically preserving morphological information and reducing semantic confusion.
Outcome:	The proposed method outperforms BPE and surpasses state-of-the-art morpheme-aware tokenization methods.

LangSAE Editing: Improving Multilingual Information Retrieval via Post-hoc Language Identity Removal (2026.acl-long)

Copied to clipboard

Dongjun Kim, Jeongho Yoon, Chanjun Park, Heuiseok Lim

Ask, Assess, and Refine: Rectifying Factual Consistency and Hallucination in LLMs with Metric-Guided Feedback Learning (2024.eacl-long)

Challenge:	Existing methods for dense retrieval in multilingual environments encode language identity alongside semantics.
Approach:	They propose a method that trains on pooled embeddings to remove language-identity signal directly in vector space.
Outcome:	The proposed method improves ranking quality and cross-language coverage across multiple languages with especially strong gains for script-distinct languages.

Copied to clipboard

Dongyub Lee, Eunhwan Park, Hodong Lee, Heuiseok Lim

Semantic Inversion, Identical Replies: Revisiting Negation Blindness in Large Language Models (2025.emnlp-main)

Challenge:	Recent advances in Large Language Models (LLMs) have heralded unprecedented capabilities in information seeking and text generation, but challenges remain regarding citation errors and generating information not present in the evidence (hallucination).
Approach:	They propose a framework to assess citation errors and hallucination using an explicit evaluation paradigm to formulate actionable natural language feedback.
Outcome:	The proposed approach improves correctness, fluency, and citation quality and reduces hallucinations in the results.

Copied to clipboard

Jinsung Kim, Seonmin Koo, Heuiseok Lim

CReTIHC: Designing Causal Reasoning Tasks about Temporal Interventions and Hallucinated Confoundings (2023.findings-emnlp)

Challenge:	Negation is a common occurrence in the real world and is essential for logical reasoning as it helps understand the opposite or absence of a statement.
Approach:	They propose a verification framework that includes task design and measurement methods to verify this phenomenon negation blindness on the query.
Outcome:	The proposed framework can be used to verify the model fails to capture semantic contradictions in negated queries despite its accurate understanding of knowledge about positive queries.

Copied to clipboard

Changwoo Chun, SongEun Lee, Jaehyung Seo, Heuiseok Lim

Don’t Judge a Language Model by Its Last Layer: Contrastive Learning with Layer-Wise Attention Pooling (2022.coling-1)

Challenge:	Large language models (LLMs) have demonstrated impressive capabilities in natural language processing, but their ability to establish causal relationships remains challenging.
Approach:	They propose a novel dataset to test and enhance the causal reasoning abilities of large language models (LLMs) by integrating elements of verbal hallucinations and temporal interventions into existing causal inference datasets.
Outcome:	The proposed dataset is designed to test and enhance the causal reasoning abilities of large language models.

Copied to clipboard

Dongsuk Oh, Yejin Kim, Hodong Lee, H. Howie Huang, Heuiseok Lim

I Know What You Asked: Graph Path Learning using AMR for Commonsense Reasoning (2020.coling-main)

Challenge:	Recent pre-trained language models (PLMs) have shown competitive performance on many natural language processing tasks.
Approach:	They propose a pooling strategy which preserves layer-wise signals captured in each layer and learns digested linguistic features for downstream tasks.
Outcome:	The proposed method improves on standard semantic textual similarity and semantic search tasks.

Copied to clipboard

Jungwoo Lim, Dongsuk Oh, Yoonna Jang, Kisu Yang, Heuiseok Lim

PEEP-Talk: A Situational Dialogue-based Chatbot for English Education (2023.acl-demo)

Challenge:	a large amount of pre-defined commonsense knowledge is available for commonsensense reasoning . humans acquire commonsence in their lives, but machines cannot learn commonseense without assistance.
Approach:	They propose an AMR-ConceptNet-Pruned (ACP) graph that is pruned from a full integrated graph . they show that the ACP graph interprets the reasoning path and predicts the correct answer .
Outcome:	The proposed graph outperforms baseline models in the commonsenseQA task . it shows that the reasoning path can be interpreted with the relations and concepts provided by the graph .

Copied to clipboard

Seungjun Lee, Yoonna Jang, Chanjun Park, Jungseob Lee, Jaehyung Seo, Hyeonseok Moon, Sugyeong Eo, Seounghoon Lee, Bernardo Yahya, Heuiseok Lim

EASE: Entity-Aware Sub-table Generation for Real-world Multi-table QA (2026.acl-long)

Challenge:	Existing chatbots lack realistic practice scenarios for English learners . existing platforms employ hand-crafted and patternmatching rules, limiting communication ability and responding appropriately to out-of-situation utterances.
Approach:	They propose a real-world situational dialogue-based chatbot for English education . it generates appropriate responses in various real-life situations while providing accurate feedback .
Outcome:	The proposed chatbot generates appropriate responses in various real-life situations while providing accurate feedback to learners.

Copied to clipboard

Myunghoon Kang, Dahyun Jung, Suhyune Son, Seonmin Koo, Changwoo Chun, Daniel Rim, Haeyoung Kwon, Yuna Hur, Heuiseok Lim

Structured Data in NLP

Challenge:	Table-based question answering (table QA) is a powerful tool for analyzing large language models.
Approach:	They propose to use noisy multi-table sets to generate sub-tables for table-based question answering.
Outcome:	The proposed framework efficiently filters out irrelevant information while incorporating pertinent table values.

Enhancing Automatic Term Extraction with Large Language Models via Syntactic Retrieval (2025.findings-acl)

Copied to clipboard

Yongchan Chun, Minhyuk Kim, Dongjun Kim, Chanjun Park, Heuiseok Lim

Syntactic Text Processing

Challenge:	Large language models (LLMs) have improved IE, but their potential for ATE has not been explored.
Approach:	They propose a retrieval-based prompting strategy that selects demonstrations according to syntactic rather than semantic similarity in a few-shot setting.
Outcome:	The proposed method improves performance on three specialized ATE benchmarks.

Cross-Lingual Optimization for Language Transfer in Large Language Models (2025.acl-long)

Copied to clipboard

Jungseob Lee, Seongtae Hong, Hyeonseok Moon, Heuiseok Lim

CoME: An Unlearning-based Approach to Conflict-free Model Editing (2025.naacl-long)

Challenge:	Adapting large language models to other languages often suffers from an overemphasis on English performance.
Approach:	They propose a cross-lingual optimization technique that efficiently transfers an English-centric LLM to a target language while preserving its English capabilities.
Outcome:	The proposed model outperforms SFT in acquiring target language proficiency and maintaining English performance in low-resource languages.

Copied to clipboard

Dahyun Jung, Jaehyung Seo, Jaewook Lee, Chanjun Park, Heuiseok Lim

KoCommonGEN v2: A Benchmark for Navigating Korean Commonsense Reasoning Challenges in Large Language Models (2024.findings-acl)

Challenge:	Large language models (LLMs) often retain outdated or incorrect information from pre-training, which undermines their reliability.
Approach:	They propose a conflict-free model editing framework that selectively removes outdated knowledge from LLMs to improve their accuracy and reliability.
Outcome:	The proposed framework improves both editing accuracy and model reliability when applied to existing editing methods.

Copied to clipboard

Jaehyung Seo, Jaewook Lee, Chanjun Park, SeongTae Hong, Seungjun Lee, Heuiseok Lim

KoLEG: On-the-Fly Korean Legal Knowledge Editing with Continuous Retrieval (2025.findings-emnlp)

Challenge:	Language models are striving to grasp commonsense reasoning, but they are lacking in Korean commons- ense benchmarks.
Approach:	They present a fine-grained benchmark dataset focused on Korean commonsense reasoning that includes multiple-choice questions across seven error categories.
Outcome:	The proposed datasets show that LLMs struggle with Korean commonsense reasoning . human accuracy benchmarked at approximately 85%, while GPT-4’s performance lags at about 74%, and other LLM models demonstrate an average accuracy of around 42%.

Copied to clipboard

Jaehyung Seo, Dahyun Jung, Jaewook Lee, Yongchan Chun, Dongjun Kim, Hwijung Ryu, Donghoon Shin, Heuiseok Lim

KEBAP: Korean Error Explainable Benchmark Dataset for ASR and Post-processing (2023.emnlp-main)

Challenge:	a recent study shows that Korean legal knowledge is subject to frequent temporal updates driven by societal needs and government policies.
Approach:	They propose a Korean Legal knowledge editing framework enhanced with continuous retrieval . they employ an Editing-Aware Learning Strategy and a LawEdit Retriever .
Outcome:	a new framework outperforms existing methods for updating legal knowledge in Korean . it maintains robust performance in sequential editing and is qualitatively validated by legal experts.

Copied to clipboard

Seonmin Koo, Chanjun Park, Jinsung Kim, Jaehyung Seo, Sugyeong Eo, Hyeonseok Moon, Heuiseok Lim

No Reader Left Behind: Multi-Agent Summaries Everyone Can Understand (2026.acl-long)

Challenge:	Conventional evaluation metrics for automatic speech recognition systems produce a singular aggregate score, which is insufficient for understanding specific system vulnerabilities.
Approach:	They propose to introduce the Korean Error Explainable Benchmark Dataset for ASR and Post-processing (KEBAP) this method enables a more balanced assessment encompassing speech recognition accuracy and user readability.
Outcome:	The proposed method enables a more balanced assessment encompassing speech recognition accuracy and user readability.

Copied to clipboard

Jimin Jung, MyoungJin Kim, Jaehyung Seo, Heuiseok Lim

Exploring Inherent Biases in LLMs within Korean Social Context: A Comparative Analysis of ChatGPT and GPT-4 (2024.naacl-srw)

Challenge:	Existing summarization systems struggle to address diverse linguistic and cognitive barriers among general readers.
Approach:	They propose a multi-agent framework that integrates template-based planning with an iterative feedback loop guided by simulated readers and domain expert revision to address comprehension barriers such as unknown terms, missing contexts, and confusing sentences.
Outcome:	The proposed framework improves readability and factuality across multiple datasets and human evaluations show that it is more accessible to a wide range of readers.

Copied to clipboard

Seungyoon Lee, Dong Kim, Dahyun Jung, Chanjun Park, Heuiseok Lim

Benchmark Profiling: Mechanistic Diagnosis of LLM Benchmarks (2025.emnlp-main)

Challenge:	Large Language Models (LLMs) have been criticized for perpetuating stereotypes against diverse groups based on race, sexual orientation, and other attributes.
Approach:	They devised a set of prompts that reflect major societal issues in Korea and assign varied personas to both ChatGPT and GPT-4 to assess the toxicity of the generated sentences.
Outcome:	The proposed model produces twice the level of toxic content as ChatGPT and GPT-4 under certain conditions.

Copied to clipboard

Dongjun Kim, Gyuho Shim, Yongchan Chun, Minhyuk Kim, Chanjun Park, Heuiseok Lim

Challenge:	Large Language Models are often judged by their scores on standard benchmarks, yet such scores often overstate real capability since they mask the mix of skills a task actually demands.
Approach:	They propose a diagnostic framework that decomposes benchmark performance into ten cognitively grounded abilities and computes an Ability Impact Score (AIS) AIS quantifies how much each ability contributes to a model’s success on a given benchmark.
Outcome:	The proposed framework decomposes performance into ten cognitively grounded abilities and computes an Ability Impact Score (AIS) that quantifies how much each ability contributes to a model’s success on a given benchmark.

Rich Character-Level Information for Korean Morphological Analysis and Part-of-Speech Tagging (C18-1)

Copied to clipboard

Andrew Matteson, Chanhee Lee, Youngbum Kim, Heuiseok Lim

Syntactic Text Processing

Challenge:	Korean is a highly agglutinative, character-rich language, requiring dictionary-less morphological analysis . a novel model can perform morphology and part-of-speech tagging without prior knowledge .
Approach:	They propose a multi-stage action-based model that performs morphological transformation and part-of-speech tagging using arbitrary units of input.
Outcome:	The proposed model achieves state-of-the-art word and sentence-level tagging accuracy with Korean corpus.

You Truly Understand What I Need : Intellectual and Friendly Dialog Agents grounding Persona and Knowledge (2022.findings-emnlp)

Copied to clipboard

Jungwoo Lim, Myunghoon Kang, Yuna Hur, Seung Won Jeong, Jinsung Kim, Yoonna Jang, Dongyub Lee, Hyesung Ji, DongHoon Shin, Seungryong Kim, Heuiseok Lim

CHEF in the Language Kitchen: A Generative Data Augmentation Leveraging Korean Morpheme Ingredients (2023.emnlp-main)

Challenge:	Existing models that ground knowledge and persona at the same time are limited, leading to hallucination and a passive way of using personas.
Approach:	They propose a conversational agent that grounds external knowledge and persona simultaneously and a retrieval augmented generation model that generates utterances with lesser hallucination and more engagingness.
Outcome:	The proposed agent generates the utterance with lesser hallucination and more engagingness utilizing retrieval augmented generation with knowledge-persona enhanced query.

Copied to clipboard

Jaehyung Seo, Hyeonseok Moon, Jaewook Lee, Sugyeong Eo, Chanjun Park, Heuiseok Lim

FLEX: A Benchmark for Evaluating Robustness of Fairness in Large Language Models (2025.findings-naacl)

Challenge:	Korean morphological variations present unique opportunities and challenges in natural language processing (NLP), necessitating an advanced understanding of morpheme-based sentence construction.
Approach:	They propose a method to replicate morphological transformations inherent in Korean sentences based on lexical and functional morphemes through generative data augmentation.
Outcome:	The proposed method improves performance in Korean multiple classification datasets without incurring external data usage.

Copied to clipboard

Dahyun Jung, Seungyoon Lee, Hyeonseok Moon, Chanjun Park, Heuiseok Lim

Priming Ancient Korean Neural Machine Translation (2022.lrec-1)

Challenge:	Existing safety evaluations may overlook the inherent weaknesses of Large Language Models, despite their benefits.
Approach:	They propose a benchmark to evaluate the robustness of Large Language Models under extreme conditions.
Outcome:	The proposed approach evaluates the fairness of large language models under extreme conditions.

Copied to clipboard

Chanjun Park, Seolhwa Lee, Jaehyung Seo, Hyeonseok Moon, Sugyeong Eo, Heuiseok Lim

Should we find another model?: Improving Neural Machine Translation Performance with ONE-Piece Tokenization Method without Model Modification (2021.naacl-industry)

Challenge:	Recent studies have focused on the restoration and translation of historical languages.
Approach:	They propose to use two different stimuli to priming ancient-Korean NMT . they confirm the possibility of developing a human-centric model based on cognitive science .
Outcome:	The proposed model can be used to translate historical Korean documents using neural machine translation.

Copied to clipboard

Chanjun Park, Sugyeong Eo, Hyeonseok Moon, Heuiseok Lim

Syntactic Text Processing

Challenge:	Recent studies using pretrain-finetuning approach have achieved state-of-the-art (SOTA) performance in many natural language processing tasks.
Approach:	They propose a new tokenization method that combines morphology-considered subword tokenization and vocabulary methods to address this limitation.
Outcome:	The proposed method can be used without modifying the model structure.

Hyper-BTS Dataset: Scalability and Enhanced Analysis of Back TranScription (BTS) for ASR Post-Processing (2024.findings-eacl)

Copied to clipboard

Chanjun Park, Jaehyung Seo, Seolhwa Lee, Junyoung Son, Hyeonseok Moon, Sugyeong Eo, Chanhee Lee, Heuiseok Lim

Exploring Coding Spot: Understanding Parametric Contributions to LLM Coding Performance (2026.findings-acl)

Challenge:	Automatic Speech Recognition (ASR) post-processing requires substantial amounts of data, requiring expensive phonetic transcription experts.
Approach:	They propose a "Hyper-BTS" dataset that is five times larger than prior studies . they propose criteria for categorizing error types within ASR post-processing .
Outcome:	The proposed method can generate ASR inputs from clean text using a text-to-speech system.

Copied to clipboard

Dongjun Kim, Minhyuk Kim, Yongchan Chun, Chanjun Park, Heuiseok Lim

Metric Calculating Benchmark: Code-Verifiable Complicate Instruction Following Benchmark for Large Language Models (2025.emnlp-main)

Challenge:	Large Language Models (LLMs) have demonstrated proficiency in code generation and comprehension across multiple programming languages.
Approach:	They propose a parameter-localized subset of LLMs that facilitates coding capabilities.
Outcome:	The proposed model significantly improves performance on coding tasks while preserving non-coding functionalities.

Copied to clipboard

Hyeonseok Moon, Seongtae Hong, Jaehyung Seo, Heuiseok Lim

Towards Precise Localization of Critical Errors in Machine Translation (2024.findings-acl)

Challenge:	Recent frontier-level LLMs have saturated many previously difficult benchmarks, leaving little room for further differentiation.
Approach:	They propose a benchmark to evaluate whether LLMs can execute string-matching NLP metrics by strictly following step-by-step instructions.
Outcome:	The proposed benchmarks show that they can perform step-by-step execution, instruction adherence, numerical computation, and long-range consistency in handling intermediate results.

Copied to clipboard

Dahyun Jung, Sugyeong Eo, Heuiseok Lim

Machine Translation

Challenge:	Existing critical error detection focuses on identifying sentence-level errors, leaving the precise localization of such errors unaddressed.
Approach:	They propose a task to detect critical errors at a fine-grained level in machine translation sentences.
Outcome:	The proposed method outperforms existing methods and LLMs in English to Korean translations.

Towards Diverse and Effective Question-Answer Pair Generation from Children Storybooks (2023.findings-acl)

Copied to clipboard

Sugyeong Eo, Hyeonseok Moon, Jinsung Kim, Yuna Hur, Jeongwook Kim, SongEun Lee, Changwoo Chun, Sungsoo Park, Heuiseok Lim

Generative Interpretation: Toward Human-Like Evaluation for Educational Question-Answer Pair Generation (2024.findings-eacl)

Challenge:	Recent advances in QA pair generation (QAG) have raised interest in applying this technique to the educational field.
Approach:	They propose a QAG framework that enhances QA type diversity by producing different interrogative sentences and implicit/explicit answers.
Outcome:	The proposed framework outperforms state-of-the-art methods by significant margins, achieving improved diversity and quality.

Copied to clipboard

Hyeonseok Moon, Jaewook Lee, Sugyeong Eo, Chanjun Park, Jaehyung Seo, Heuiseok Lim

Explainability & Interpretability in NLP

Challenge:	Existing evaluation methods often fail to produce objective results and favor high similarity to the ground-truth question-answer pairs.
Approach:	They propose an alternative approach to evaluate question-answer generation using Generative Interpretation (GI) GI outperforms existing evaluation methods in terms of human alignment .
Outcome:	The proposed approach outperforms existing evaluation methods in human alignment and shows comparable performance with GPT3.5, only with BART-large.

CharacterGPT: A Persona Reconstruction Framework for Role-Playing Agents (2025.naacl-industry)

Copied to clipboard

Jeiyoon Park, Chanjun Park, Heuiseok Lim

KoCHET: A Korean Cultural Heritage Corpus for Entity-related Tasks (2022.coling-1)

Challenge:	Maintaining consistent character personas remains a significant challenge due to variability in information extraction.
Approach:	They propose a framework to dynamically reconstruct character personas through Character Persona Training.
Outcome:	The proposed framework is evaluated through Big Five personality evaluations and creative tasks, in which characters generate original narratives.

Copied to clipboard

Gyeongmin Kim, Jinsung Kim, Junyoung Son, Heuiseok Lim

Unveiling the Limits of Large Language Models in Inferring Pragmatic Meaning from Non-Verbal Responses (2026.acl-long)

Challenge:	Existing corpus for entity-related tasks is limited in terms of application and cannot be used for entity recognition.
Approach:	They propose to use a Korean cultural heritage corpus for the typical entity-related tasks named entity recognition (NER), relation extraction (RE) and entity typing (ET) .
Outcome:	The proposed corpus makes it more useful in terms of cultural heritage and provides practical insights in terms linguistic analysis.

Copied to clipboard

Sugyeong Eo, Heuiseok Lim

PANDA: Persona Attributes Navigation for Detecting and Alleviating Overuse Problem in Large Language Models (2024.emnlp-main)

Challenge:	Existing studies have focused mainly on LLMs' comprehension of verbal behavior, with non-verbal behavior considered only in conjunction with verbal responses.
Approach:	They present the first systematic evaluation of LLMs’ ability to infer pragmatic meaning in dialogue consisting solely of non-verbal responses.
Outcome:	The proposed model fails to capture non-verbal intent and has accuracy dropping by 60% compared to verbal ones.

Copied to clipboard

Jinsung Kim, Seonmin Koo, Heuiseok Lim

Dialogue Systems & Conversational Agents

Challenge:	In persona-grounded dialogue, it is required to respond fluently and ground attributes according to the current conversation topic properly.
Approach:	They propose a framework to quantify the persona overuse problem of LLMs by establishing clear standards and verifying various LLM based on them.
Outcome:	The proposed framework aims to quantify the persona overuse problem of LLMs by establishing clear standards and verifying various LLM based on them.

Reliable Evaluation Protocol for Low-Precision Retrieval (2026.acl-short)

Copied to clipboard

Kisu Yang, Yoonna Jang, Hwanseok Jang, Kenneth Choi, Isabelle Augenstein, Heuiseok Lim

I Don’t Need Solution. I Need Emotional Support : Empathetic LLMs based on Emotional Validation (2026.findings-acl)

Challenge:	Recent studies have shown that low-precision methods can improve performance, but they introduce high variability in the results based on tie resolution.
Approach:	They propose a retrieval evaluation protocol designed to reduce tie variation . high-precision scoring and tie-aware retrieval metrics are proposed to reduce this variability .
Outcome:	The proposed retrieval evaluation protocol reduces tie-induced instability and recovers expected scores and ranges on 12 retrieval datasets.

Copied to clipboard

Suhyune Son, Jungwoo Lim, Myunghoon Kang, Seongtae Hong, Yuna Hur, Evelyn H. Zi, Heuiseok Lim

Post-hoc Utterance Refining Method by Entity Mining for Faithful Knowledge Grounded Conversations (2023.emnlp-main)

Challenge:	Existing large language models (LLMs) struggle to generate emotional support response, despite observing and reflecting on the help-seeker’s situation . Empathy drives the formation of constructive interpersonal and supportive relationships, including counseling for mental health care .
Approach:	They propose to use a two-stage training process to enhance empathetic response generation through empathy acquisition and emotional validation alignment.
Outcome:	The proposed method significantly improves empathetic response generation, achieving superior performance in both automatic and human evaluations.

Copied to clipboard

Yoonna Jang, Suhyune Son, Jeongwoo Lee, Junyoung Son, Yuna Hur, Jungwoo Lim, Hyeonseok Moon, Kisu Yang, Heuiseok Lim

Towards Scalable Lifelong Knowledge Editing with Selective Knowledge Suppression (2026.acl-long)

Challenge:	Despite advances in language generation, models suffer from hallucinations that are either untrue or unfaithful to a given source.
Approach:	They propose a method to refine hallucinated utterances based on source knowledge . REM implicitly uses key entities in the knowledge to refine the utterant .
Outcome:	The proposed method reduces entity hallucination in the generated utterance and improves the quality of the model.

Copied to clipboard

Dahyun Jung, Jaewook Lee, Heuiseok Lim

Explainable CED: A Dataset for Explainable Critical Error Detection in Machine Translation (2024.naacl-srw)

Challenge:	Existing methods to modify knowledge are limited due to high training costs and lack stability during sequential edits due to catastrophic forgetting.
Approach:	They propose a framework to modify specific knowledge of large language models without retraining the entire model.
Outcome:	Extensive experiments on ZSRE, Counterfact, and RIPE show that LightEdit outperforms existing lifelong knowledge editing methods.

Copied to clipboard

Dahyun Jung, Sugyeong Eo, Chanjun Park, Heuiseok Lim

Explainability & Interpretability in NLP

Challenge:	Existing studies of critical error detection lack content addressing the causes of catastrophic errors.
Approach:	They propose a dataset that introduces the attributes of error explanation and correction regarding critical errors.
Outcome:	The proposed dataset reduces time costs and mitigates human annotation bias.

MIRAGE: A Metric-Intensive Benchmark for Retrieval-Augmented Generation Evaluation (2025.findings-naacl)

Copied to clipboard

Chanhee Park, Hyeonseok Moon, Chanjun Park, Heuiseok Lim

Empirical Analysis of Noising Scheme based Synthetic Data Generation for Automatic Post-editing (2022.lrec-1)

Challenge:	Retrieval-Augmented Generation (RAG) systems are limited in their evaluation due to the intricate interplay between retrieval and generation components.
Approach:	They propose a Question Answering Question Answerer dataset specifically designed for RAG evaluation that integrates external, non-parametric knowledge retrieved by a retrieval pool of 37,800 entries.
Outcome:	The proposed dataset consists of 7,560 curated instances mapped to a retrieval pool of 37,800 entries, enabling an efficient evaluation of both retrieval and generation tasks.

Copied to clipboard

Hyeonseok Moon, Chanjun Park, Seolhwa Lee, Jaehyung Seo, Jungseob Lee, Sugyeong Eo, Heuiseok Lim

Find the Intention of Instruction: Comprehensive Evaluation of Instruction Understanding for Large Language Models (2025.findings-naacl)

Challenge:	Automatic post-editing (APE) is a research field that aims to correct errors in translated sentences regardless of the utilized machine translation system.
Approach:	They propose a method for automatically generating APE data based on a noising scheme from a parallel corpus.
Outcome:	The proposed method shows that depending on the type of noise, the noising scheme-based APE data generation may lead to inferior performance.

Copied to clipboard

Hyeonseok Moon, Jaehyung Seo, Seungyoon Lee, Chanjun Park, Heuiseok Lim

FreeTalky: Don’t Be Afraid! Conversations Made Easier by a Humanoid Robot using Persona-based Dialogue (2022.lrec-1)

Challenge:	LLMs are prone to generate responses to instruction-formatted statements in an instinctive manner, rather than comprehending the underlying user intention within the given instructions.
Approach:	They propose to use an instruction-following capability benchmark to evaluate LLMs' instruction understanding capability.
Outcome:	The proposed benchmark analyzes the instruction understanding capability of large language models with four instruction candidates and a single candidate.

Copied to clipboard

Chanjun Park, Yoonna Jang, Seolhwa Lee, Sungjin Park, Heuiseok Lim

Translation of Multifaceted Data without Re-Training of Machine Translation Systems (2024.findings-emnlp)

Challenge:	FreeTalky is a deep learning-based foreign language learning platform for people who experience anxiety dealing with foreign languages.
Approach:	They propose a deep learning-based foreign language learning platform called FreeTalky . it employs a humanoid robot NAO and various deep learning models .
Outcome:	The proposed system provides personalized learning based on persona dialogue and grammar error correction, and also helps alleviate xenoglossophobia by replacing the real human in the conversation with a NAO robot, through human evaluation.

Copied to clipboard

Hyeonseok Moon, Seungyoon Lee, SeongTae Hong, Seungjun Lee, Chanjun Park, Heuiseok Lim

Character-Level Feature Extraction with Densely Connected Networks (C18-1)

Challenge:	a novel MT pipeline that considers the intra-data relation is proposed . previous MT systems have demonstrated relatively low performance, making them hardly utilized as another data source.
Approach:	They propose a new MT pipeline that considers the intra-data relation . they propose CS and IT to enhance the intra data relation based on a data point .
Outcome:	The proposed pipeline improves translation quality and training data compared with the existing approach . it yields better training data and better translation quality than previous approaches .

Copied to clipboard

Chanhee Lee, Young-Bum Kim, Dongyub Lee, Heuiseok Lim

Beyond Candidates : Adaptive Dialogue Agent Utilizing Persona and Knowledge (2023.findings-emnlp)

Challenge:	Existing methods to generate character-level features with neural architectures such as CNN or Recurrent Neural Network (RNN) are slow and generate position-independent features.
Approach:	They propose a method that uses a densely connected network to extract character-level features from words using CNN and RNN.
Outcome:	The proposed method shows robustness and effectiveness while being faster than CNN- or RNN-based methods.

Copied to clipboard

Jungwoo Lim, Myunghoon Kang, Jinsung Kim, Jeongwook Kim, Yuna Hur, Heuiseok Lim

Dialogue Systems & Conversational Agents

Challenge:	a previous study suggested that human dialogue systems ground persona and knowledge but they require incomplete candidate sets.
Approach:	They propose an adaptive dialogue agent that uses persona and knowledge without candidate sets . their model generates consistent and relevant persona descriptions and identifies relevant knowledge .
Outcome:	The proposed model outperforms baselines that ground persona and knowledge candidates even with fragmentary information.

Towards Privacy-Preserving Large Language Model: Text-free Inference Through Alignment and Adaptation (2026.acl-long)

Copied to clipboard

Jeongho Yoon, Chanhee Park, Yongchan Chun, Hyeonseok Moon, Heuiseok Lim

PicTalky: Augmentative and Alternative Communication for Language Developmental Disabilities (2022.aacl-demo)

Challenge:	Existing LLMs require users to submit raw text regardless of its sensitivity, resulting in substantial computational overhead and degrade model performance.
Approach:	They propose a new training pipeline that allows a client-side encoder to condition on k-pooled prompt embeddings instead of raw text and a server-side projection module to fine-tune the projection module and LLM on private, domain-specific data using noise-injected embeddables.
Outcome:	The proposed approach eliminates the need for transmitting raw prompt text while maintaining a favorable balance between privacy preservation and model utility for both clients and service providers.

Copied to clipboard

Chanjun Park, Yoonna Jang, Seolhwa Lee, Jaehyung Seo, Kisu Yang, Heuiseok Lim

Challenge:	Existing software packages are expensive and difficult to use, and only provide simple functions.
Approach:	They propose an AI-based AAC system called PicTalky that can improve communication skills for children with language disabilities.
Outcome:	The proposed system improves communication skills and language comprehension abilities for children with language disabilities.

CLEAR: Cross-Lingual Enhancement in Retrieval via Reverse-training (2026.acl-long)

Copied to clipboard

Seungyoon Lee, Minhyuk Kim, Seongtae Hong, Youngjoon Jang, Dongsuk Oh, Heuiseok Lim

HAWK: Highlighting Entity-aware Knowledge for Alleviating Information Sparsity in Long Contexts (2025.findings-emnlp)

Challenge:	Existing multilingual embedding models often struggle to capture cross-lingual alignment during training.
Approach:	They propose a novel loss function that leverages an English passage as a bridge to strengthen alignments between target language and English.
Outcome:	The proposed model improves retrieval performance across cross-lingual scenarios while minimizing performance degradation in English.

Copied to clipboard

Seonmin Koo, Jinsung Kim, Chanjun Park, Heuiseok Lim

Question Answering

Challenge:	a problem of information sparsity in QA tasks is causing fragmentation of textual data . highlighting entity-AWare Knowledge (HAWK) framework can be used to address this problem .
Approach:	a framework is proposed to highlight key information in a context and structuralize it in an entity-aware manner.
Outcome:	a proposed framework improves QA tasks with long contexts by highlighting key information in a context . the framework achieves a 27.6-point F1 score increase and an average win rate of 76.75% .

Mixture-of-Clustered-Experts: Advancing Expert Specialization and Generalization in Instruction Tuning (2025.emnlp-main)

Copied to clipboard

Sugyeong Eo, Jung Jun Lee, Chanjun Park, Heuiseok Lim

Text Clustering

Challenge:	A sparse Mixture-of-Experts architecture has emerged as a highly scalable solution for instruction tuning.
Approach:	They propose a mixture-of-Clustered-Experts (MoCE) architecture that allows expert specialization . they evaluate the mechanism on a set of benchmarks and show its superiority .
Outcome:	The proposed approach outperforms existing models and benchmarks on instruction tuning scenarios with significant input heterogeneity.

Leveraging Pre-existing Resources for Data-Efficient Counter-Narrative Generation in Korean (2024.lrec-main)

Copied to clipboard

Seungyoon Lee, Chanjun Park, DaHyun Jung, Hyeonseok Moon, Jaehyung Seo, Sugyeong Eo, Heuiseok Lim

NeedleChain: Measuring Intact Context Comprehension Capability of Large Language Models (2026.findings-acl)

Challenge:	Existing datasets and methods for detecting hate speech are limited by resource-intensive nature and only focus on the primary language.
Approach:	They propose a Korean Hate Speech Counter Punch (KHSCP) method that generates fact-based responses to hate speech in the Korean language and propose to use existing resources to overcome data scarcity.
Outcome:	The proposed method can overcome data scarcity in low-resource environments by leveraging existing resources.

Copied to clipboard

Hyeonseok Moon, Heuiseok Lim

GRASP: Guiding Model with RelAtional Semantics Using Prompt for Dialogue Relation Extraction (2022.coling-1)

Challenge:	Existing benchmarks for context understanding embed query-irrelevant content . this shifts evaluation toward retrieving relevant snippets rather than fully integrating all provided information.
Approach:	They propose a benchmark to evaluate whether models can faithfully incorporate all given evidence . they propose 'needlechain' benchmark to test whether models incorporate all available information .
Outcome:	The proposed benchmarks overestimate the ability of large language models to integrate all given evidence when the context is entirely query-relevant.

Copied to clipboard

Junyoung Son, Jinsung Kim, Jungwoo Lim, Heuiseok Lim

REVISE: A Framework for Revising OCRed text in Practical Information Systems with Data Contamination Strategy (2025.acl-industry)

Challenge:	Existing studies utilize pre-trained language models with extensive features to supplement the low information density of the dialogue by multiple speakers.
Approach:	They propose a dialogue-based relation extraction task that leverages pre-trained language models to capture relational semantic clues of a given dialogue using an argument-aware prompt marker strategy and a relational clue detection task.
Outcome:	The proposed model achieves state-of-the-art on a DialogRE dataset even though it only leverages pre-trained language models without adding any extra layers.

Copied to clipboard

Gyuho Shim, Seongtae Hong, Heuiseok Lim

Visual Data in NLP

Challenge:	Existing Document AI frameworks lack the capability to structurally organize and manage document information.
Approach:	They propose a framework that corrects OCR errors at the character, word, and structural levels and a synthetic data generation strategy that realistically simulates such errors to train an effective correction model.
Outcome:	The proposed framework improves document retrieval and question answering tasks by correcting errors introduced by OCR errors at the character, word, and structural levels.

HiKEY: Hierarchical Multimodal Retrieval for Open-Domain Document Question Answering (2026.acl-long)

Copied to clipboard

Joongmin Shin, Gyuho Shim, Jeongbae Park, Jaehyung Seo, Heuiseok Lim

The Impact of Negated Text on Hallucination with Large Language Models (2025.emnlp-main)

Challenge:	Existing approaches to document-based Opendomain Question Answering (ODQA) use flat text chunks or page-level images to locate the correct document.
Approach:	They propose a hierarchical tree-based multimodal retrieval framework that elevates document hierarchy to a first-class retrieval signal.
Outcome:	The proposed framework outperforms page- and chunk-based baselines on ODQA benchmarks and improves retrieval recall by 12.9% and end-to-end QA performance by 6.8%.

Copied to clipboard

Jaehyung Seo, Hyeonseok Moon, Heuiseok Lim

From Ambiguity to Accuracy: The Transformative Effect of Coreference Resolution on Retrieval-Augmented Generation systems (2025.acl-srw)

Challenge:	Recent studies on hallucination in large language models (LLMs) have been actively progressing in natural language processing.
Approach:	They propose to examine whether LLMs can recognize contextual shifts caused by negation and still reliably distinguish hallucinations comparable to affirmative cases.
Outcome:	The proposed model can detect hallucinations comparable to affirmative cases, but it is difficult to detect them in negated text, the authors show .

Copied to clipboard

Youngjoon Jang, Seongtae Hong, Junyoung Son, Sungjin Park, Chanjun Park, Heuiseok Lim

Explore the Way: Exploring Reasoning Path by Bridging Entities for Effective Cross-Document Relation Extraction (2023.findings-emnlp)

Challenge:	Retrieval-augmented generation (RAG) is a key framework in natural language processing . however, the effectiveness of RAG is often hindered by coreferential complexity in retrieved documents .
Approach:	They investigate how entity coreference affects document retrieval and generative performance in RAG-based systems.
Outcome:	The proposed model improves QA performance and retrieval relevance and contextual understanding.

Copied to clipboard

Junyoung Son, Jinsung Kim, Jungwoo Lim, Yoonna Jang, Heuiseok Lim

TORSO: Template-Oriented Reasoning Towards General Tasks (2025.emnlp-main)

Challenge:	Existing studies focus on capturing implicit relations between entities, but humans use explicit information chains to find the relations between two entities.
Approach:	They propose a cross-document relation extraction task that captures the relation between two entities mentioned in different documents.
Outcome:	The proposed model outperforms baseline models in the CodRED task.

Copied to clipboard

Minhyuk Kim, Seungyoon Lee, Heuiseok Lim

Low-Resource NLP

Challenge:	Existing approaches to generate responses using few-shot examples depend on the provided examples, limiting the model’s reasoning capabilities.
Approach:	They propose a model that emulates human reasoning during response generation by using curated few-shot prompts instead of manually crafted few-shot examples.
Outcome:	The proposed model achieves strong performance on diverse LLMs benchmarks with reasonable rationales.

MIGRATE: Cross-Lingual Adaptation of Domain-Specific LLMs through Code-Switching and Embedding Transfer (2025.coling-main)

Copied to clipboard

Seongtae Hong, Seungyoon Lee, Hyeonseok Moon, Heuiseok Lim

Search if you don’t know! Knowledge-Augmented Korean Grammatical Error Correction with Large Language Models (2024.findings-emnlp)

Challenge:	Large Language Models (LLMs) have advanced in many fields, but focus on English-centric models requires extensive data.
Approach:	They propose a method that leverages open-source static embedding models and up to 3 million tokens of code-switching data to facilitate the seamless transfer of embeddables to target languages.
Outcome:	The proposed method outperforms baseline and existing cross-lingual transfer methods in target languages.

Copied to clipboard

Seonmin Koo, Jinsung Kim, Chanjun Park, Heuiseok Lim

QUAK: A Synthetic Quality Estimation Dataset for Korean-English Neural Machine Translation (2022.coling-1)

Challenge:	Existing studies have shown that the performance of large language models is insufficient for non-English data, such as Korean.
Approach:	They propose a framework that integrates evidential information from external sources into the prompt for the Korean GEC task.
Outcome:	The proposed framework extracts salient phrases from the given source and retrieves non-parametric knowledge based on these phrases.

Copied to clipboard

Sugyeong Eo, Chanjun Park, Hyeonseok Moon, Jaehyung Seo, Gyeongmin Kim, Jungseob Lee, Heuiseok Lim

Intelligent Predictive Maintenance RAG framework for Power Plants: Enhancing QA with StyleDFS and Domain Specific Instruction Tuning (2024.emnlp-industry)

Challenge:	despite its high utility, there are limitations concerning manual QE data creation.
Approach:	They propose to generate a Korean-English QE dataset that is fully automatic . they find that the algorithm is more accurate and faster than manual QE .
Outcome:	The proposed datasets show that they scale up to 1.58M and 6.58M, respectively, and show that the results are significantly better when compared to the previous datasets.

Copied to clipboard

Seongtae Hong, Joong Shin, Jaehyung Seo, Taemin Lee, Jeongbae Park, Cho Young, Byeongho Choi, Heuiseok Lim