GLGary Geunbae Lee

Gary Geunbae Lee

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Gary Geunbae Lee has a total of 98 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    S 3 : Sequential Self-feeding Slot Prediction for Explainable Cognitive Reframing in Task-Oriented Dialogue

    by Subin Kim, Hoonrae Kim, Gary Geunbae Lee

    Cognitive behavioral therapy (CBT) is a well-established psychotherapeutic approach, yet its delivery through AI systems remains challenging: small language models (SLMs) struggle with multi-step therapeutic reasoning, while large language models (LLMs) are often impractical for resource-constrained deployment. We address this gap by formulating CBT-based cognitive reframing as a task-oriented dialogue (TOD) problem with dialogue state tracking (DST), which exposes intermediate belief states—such as event, thought, and cognitive distortion—making the reasoning process explainable and auditable for clinical oversight. To support controlled experimentation, we introduce CREPAN , a synthetic dataset of over 6,800 expert-guided CBT reframing dialogues for panic disorder, designed as a controlled experimental setting to investigate whether models can learn the structured reasoning prescribed by the CBT ABC model. We further propose Sequential Self-feeding Slot Prediction ( \(\mbox{S}^{3}\) ), a lightweight reasoning scaffold that guides models through interpretable cognitive stages in a theoretically grounded order derived from Ellis’s ABC model. Experiments show that \(\mbox{S}^{3}\) -T5 attains a JGA of 0.7781, narrowing the gap with 8B-scale LLMs, with the sequential self-feeding structure enabling interpretable step-by-step reasoning. These results suggest that cognitive-theoretic structure is central to robust therapeutic state tracking, offering a practical path toward lightweight, interpretable, and clinically accountable AI systems for CBT-based counseling (Code and data are available at https://github.com/nobel-postech/s3-tod ).

  • Open Access

    Mixture-of-Experts with Intermediate CTC Supervision for Accented Speech Recognition

    by Wonjun Lee, Hyounghun Kim, Gary Geunbae Lee

    Accented speech remains a persistent challenge for automatic speech recognition (ASR), as most models are trained on data dominated by a few high-resource English varieties, leading to substantial performance degradation for other accents. Accent-agnostic approaches improve robustness yet struggle with heavily accented or unseen varieties, while accent-specific methods rely on limited and often noisy labels. We introduce Moe-Ctc, a Mixture-of-Experts architecture with intermediate CTC supervision that jointly promotes expert specialization and generalization. During training, accent-aware routing encourages experts to capture accent-specific patterns, which gradually transitions to label-free routing for inference. Each expert is equipped with its own CTC head to align routing with transcription quality, and a routing-augmented loss further stabilizes optimization. Experiments on the Mcv-Accent benchmark demonstrate consistent gains across both seen and unseen accents in low- and high-resource conditions, achieving up to 29.3% relative WER reduction over strong FastConformer baselines.

  • Open Access

    Difficulty-Controllable Cloze Question Distractor Generation

    by Seokhoon Kang, Yejin Jeon, Seonjeong Hwang, Gary Geunbae Lee

    Multiple-choice cloze questions are commonly used to assess linguistic proficiency and comprehension. However, generating high-quality distractors remains challenging, as existing methods often lack adaptability and control over difficulty levels, and the absence of difficulty-annotated datasets further hinders progress. To address these issues, we propose a novel framework for generating distractors with controllable difficulty by leveraging both data augmentation and a multitask learning strategy. First, to create a high-quality, difficulty-annotated dataset, we introduce a two-way distractor generation process in order to produce diverse and plausible distractors. These candidates are subsequently refined through filtering and then categorized by difficulty using an ensemble QA system. Second, this newly created dataset is leveraged to train a difficulty-controllable generation model via multitask learning. The framework includes carefully designed auxiliary tasks that enhance the model's semantic understanding of distractors and its ability to estimate their difficulty. Experimental results demonstrate that our method generates high-quality distractors across difficulty levels and substantially outperforms GPT-4o in aligning distractor difficulty with human perception.

  • Open Access

    Can LLMs Estimate Cognitive Complexity of Reading Comprehension Items?

    by Seonjeong Hwang, Hyounghun Kim, Gary Geunbae Lee

    Estimating the cognitive complexity of reading comprehension (RC) items is crucial for assessing item difficulty before it is administered to learners. Unlike syntactic and semantic features, such as passage length or semantic similarity between options, cognitive features that arise during answer reasoning are not readily extractable using existing NLP tools and have traditionally relied on human annotation. In this study, we examine whether large language models (LLMs) can estimate the cognitive complexity of RC items by focusing on two dimensions-Evidence Scope and Transformation Level-that indicate the degree of cognitive burden involved in reasoning about the answer. Our experimental results demonstrate that LLMs can approximate the cognitive complexity of items, indicating their potential as tools for prior difficulty analysis. Further analysis reveals a gap between LLMs' reasoning ability and their metacognitive awareness: even when they produce correct answers, they sometimes fail to correctly identify the features underlying their own reasoning process.

  • Open Access

    KoBLEX: Open Legal Question Answering with Multi-hop Reasoning

    by Jihyung Lee, Daehui Kim, Seonjeong Hwang, Hyounghun Kim, Gary Lee

    Large Language Models (LLM) have achieved remarkable performances in general domains and are now extending into the expert domain of law. Several benchmarks have been proposed to evaluate LLMs' legal capabilities. However, these benchmarks fail to evaluate open-ended and provision-grounded Question Answering (QA). To address this, we introduce a Korean Benchmark for Legal EXplainable QA (KoBLEX), designed to evaluate provision-grounded, multi-hop legal reasoning. KoBLEX includes 226 scenario-based QA instances and their supporting provisions, created using a hybrid LLM-human expert pipeline. We also propose a method called Parametric provision-guided Selection Retrieval (ParSeR), which uses LLM-generated parametric provisions to guide legally grounded and reliable answers. ParSeR facilitates multi-hop reasoning on complex legal questions by generating parametric provisions and employing a three-stage sequential retrieval process. Furthermore, to better evaluate the legal fidelity of the generated answers, we propose Legal Fidelity Evaluation (LF-Eval). LF-Eval is an automatic metric that jointly considers the question, answer, and supporting provisions and shows a high correlation with human judgments. Experimental results show that ParSeR consistently outperforms strong baselines, achieving the best results across multiple LLMs. Notably, compared to standard retrieval with GPT-4o, ParSeR achieves +37.91 higher F1 and +30.81 higher LF-Eval. Further analyses reveal that ParSeR efficiently delivers consistent performance across reasoning depths, with ablations confirming the effectiveness of ParSeR.

  • Open Access

    Safeguarding RAG Pipelines with GMTP: A Gradient-based Masked Token Probability Method for Poisoned Document Detection

    by San Kim, Jonghwi Kim, Yejin Jeon, Gary Geunbae Lee

    Retrieval-Augmented Generation (RAG) enhances Large Language Models (LLMs) by providing external knowledge for accurate and up-to-date responses. However, this reliance on external sources exposes a security risk, attackers can inject poisoned documents into the knowledge base to steer the generation process toward harmful or misleading outputs. In this paper, we propose Gradient-based Masked Token Probability (GMTP), a novel defense method to detect and filter out adversarially crafted documents. Specifically, GMTP identifies high-impact tokens by examining gradients of the retriever's similarity function. These key tokens are then masked, and their probabilities are checked via a Masked Language Model (MLM). Since injected tokens typically exhibit markedly low masked-token probabilities, this enables GMTP to easily detect malicious documents and achieve high-precision filtering. Experiments demonstrate that GMTP is able to eliminate over 90% of poisoned content while retaining relevant documents, thus maintaining robust retrieval and generation performance across diverse datasets and adversarial settings.

  • Open Access

    DeRAGEC: Denoising Named Entity Candidates with Synthetic Rationale for ASR Error Correction

    by Solee Im, Wonjun Lee, Jinmyeong An, Yunsu Kim, Jungseul Ok, Gary Geunbae Lee

    We present DeRAGEC, a method for improving Named Entity (NE) correction in Automatic Speech Recognition (ASR) systems. By extending the Retrieval-Augmented Generative Error Correction (RAGEC) framework, DeRAGEC employs synthetic denoising rationales to filter out noisy NE candidates before correction. By leveraging phonetic similarity and augmented definitions, it refines noisy retrieved NEs using in-context learning, requiring no additional training. Experimental results on CommonVoice and STOP datasets show significant improvements in Word Error Rate (WER) and NE hit ratio, outperforming baseline ASR and RAGEC methods. Specifically, we achieved a 28% relative reduction in WER compared to ASR without postprocessing. Our source code is publicly available at: https://github.com/solee0022/deragec

  • Open Access

    EnSToM: Enhancing Dialogue Systems with Entropy-Scaled Steering Vectors for Topic Maintenance

    by Heejae Suh, Yejin Jeon, Deokhyung Kang, Taehee Park, Yejin Min, Gary Geunbae Lee

    Small large language models (sLLMs) offer the advantage of being lightweight and efficient, which makes them suitable for resource-constrained environments. However, sLLMs often struggle to maintain topic consistency in task-oriented dialogue systems, which is critical for scenarios such as service chatbots. Specifically, it is important to ensure that the model denies off-topic or malicious inputs and adheres to its intended functionality so as to prevent potential misuse and uphold reliability. Towards this, existing activation engineering approaches have been proposed to manipulate internal activations during inference. While these methods are effective in certain scenarios, our preliminary experiments reveal their limitations in ensuring topic adherence. Therefore, to address this, we propose a novel approach termed Entropy-scaled Steering vectors for Topic Maintenance (EnSToM). EnSToM dynamically adjusts the steering intensity based on input uncertainty, which allows the model to handle off-topic distractors effectively while preserving on-topic accuracy. Our experiments demonstrate that EnSToM achieves significant performance gain with a relatively small data size compared to fine-tuning approaches. By improving topic adherence without compromising efficiency, our approach provides a robust solution for enhancing sLLM-based dialogue systems.

  • Open Access

    PicPersona-TOD : A Dataset for Personalizing Utterance Style in Task-Oriented Dialogue with Image Persona

    by Jihyun Lee, Yejin Jeon, Seungyeon Seo, Gary Geunbae Lee

    Task-Oriented Dialogue (TOD) systems are designed to fulfill user requests through natural language interactions, yet existing systems often produce generic, monotonic responses that lack individuality and fail to adapt to users' personal attributes. To address this, we introduce PicPersona-TOD, a novel dataset that incorporates user images as part of the persona, enabling personalized responses tailored to user-specific factors such as age or emotional context. This is facilitated by first impressions, dialogue policy-guided prompting, and the use of external knowledge to reduce hallucinations. Human evaluations confirm that our dataset enhances user experience, with personalized responses contributing to a more engaging interaction. Additionally, we introduce a new NLG model, Pictor, which not only personalizes responses, but also demonstrates robust performance across unseen domains https://github.com/JihyunLee1/PicPersona.

  • Open Access

    Revisiting Early Detection of Sexual Predators via Turn-level Optimization

    by Jinmyeong An, Sangwon Ryu, Heejin Do, Yunsu Kim, Jungseul Ok, Gary Geunbae Lee

    Online grooming is a severe social threat where sexual predators gradually entrap child victims with subtle and gradual manipulation. Therefore, timely intervention for online grooming is critical for proactive protection. However, previous methods fail to determine the optimal intervention points (i.e., jump to conclusions) as they rely on chat-level risk labels by causing weak supervision of risky utterances. For timely detection, we propose speed control reinforcement learning (SCoRL) (The code and supplementary materials are available at https://github.com/jinmyeongAN/SCoRL), incorporating a practical strategy derived from luring communication theory (LCT). To capture the predator's turn-level entrapment, we use a turn-level risk label based on the LCT. Then, we design a novel speed control reward function that balances the trade-off between speed and accuracy based on turn-level risk label; thus, SCoRL can identify the optimal intervention moment. In addition, we introduce a turn-level metric for precise evaluation, identifying limitations in previously used chat-level metrics. Experimental results show that SCoRL effectively preempted online grooming, offering a more proactive and timely solution. Further analysis reveals that our method enhances performance while intuitively identifying optimal early intervention points.