Publications (75)
Out-of-Manifold Regularization in Contextual Embedding Space for Text Classification
Seonghyeon Lee, Dongha Lee, Hwanjo Yu
Recent studies on neural networks with pre-trained weights (i.e., BERT) have mainly focused on a low-dimensional subspace, where the embedding vectors computed from input words (or…
Topic Coverage-based Demonstration Retrieval for In-Context Learning
Wonbin Kweon, SeongKu Kang, Runchu Tian +3
The effectiveness of in-context learning relies heavily on selecting demonstrations that provide all the necessary information for a given test input. To achieve this, it is crucia…
Topic Taxonomy Expansion via Hierarchy-Aware Topic Phrase Generation
Dongha Lee, Jiaming Shen, Seonghyeon Lee +3
Topic taxonomies display hierarchical topic structures of a text corpus and provide topical knowledge to enhance various NLP applications. To dynamically incorporate new topic info…
TaxoCom: Topic Taxonomy Completion with Hierarchical Discovery of Novel Topic Clusters
Dongha Lee, Jiaming Shen, SeongKu Kang +3
Topic taxonomies, which represent the latent topic (or category) structure of document collections, provide valuable knowledge of contents in many applications such as web search a…
Sparse Network Inversion for Key Instance Detection in Multiple Instance Learning
Beomjo Shin, Junsu Cho, Hwanjo Yu +1
Multiple Instance Learning (MIL) involves predicting a single label for a bag of instances, given positive or negative labels at bag-level, without accessing to label for each inst…
Everyday Physics in Korean Contexts: A Culturally Grounded Physical Reasoning Benchmark
Jihae Jeong, DaeYeop Lee, DongGeon Lee +1
Existing physical commonsense reasoning benchmarks predominantly focus on Western contexts, overlooking cultural variations in physical problem-solving. To address this gap, we int…
Distillation from Heterogeneous Models for Top-K Recommendation
SeongKu Kang, Wonbin Kweon, Dongha Lee +3
Recent recommender systems have shown remarkable performance by using an ensemble of heterogeneous models. However, it is exceedingly costly because it requires resources and infer…
Doubly Calibrated Estimator for Recommendation on Data Missing Not At Random
Wonbin Kweon, Hwanjo Yu
Recommender systems often suffer from selection bias as users tend to rate their preferred items. The datasets collected under such conditions exhibit entries missing not at random…
Learning Heterogeneous Temporal Patterns of User Preference for Timely Recommendation
Junsu Cho, Dongmin Hyun, SeongKu Kang +1
Recommender systems have achieved great success in modeling user's preferences on items and predicting the next item the user would consume. Recently, there have been many efforts…
Dynamic Multi-period Experts for Online Time Series Forecasting
Seungha Hong, Sukang Chae, Suyeon Kim +2
Online Time Series Forecasting (OTSF) requires models to continuously adapt to concept drift. However, existing methods often treat concept drift as a monolithic phenomenon. To add…
Improving Retrieval in Theme-specific Applications using a Corpus Topical Taxonomy
SeongKu Kang, Shivam Agarwal, Bowen Jin +3
Document retrieval has greatly benefited from the advancements of large-scale pre-trained language models (PLMs). However, their effectiveness is often limited in theme-specific ap…
Federated Continual Recommendation
Jaehyung Lim, Wonbin Kweon, Woojoo Kim +4
The increasing emphasis on privacy in recommendation systems has led to the adoption of Federated Learning (FL) as a privacy-preserving solution, enabling collaborative training wi…
Neural Sentence Embedding using Only In-domain Sentences for Out-of-domain Sentence Detection in Dialog Systems
Seonghan Ryu, Seokhwan Kim, Junhwi Choi +2
To ensure satisfactory user experience, dialog systems must be able to determine whether an input sentence is in-domain (ID) or out-of-domain (OOD). We assume that only ID sentence…
Are Vision-Language Models Safe in the Wild? A Meme-Based Benchmark Study
DongGeon Lee, Joonwon Jang, Jihae Jeong +1
Rapid deployment of vision-language models (VLMs) magnifies safety risks, yet most evaluations rely on artificial images. This study asks: How safe are current VLMs when confronted…
Learnable Dynamic Temporal Pooling for Time Series Classification
Dongha Lee, Seonghyeon Lee, Hwanjo Yu
With the increase of available time series data, predicting their class labels has been one of the most important challenges in a wide range of disciplines. Recent studies on time…
Valid Necessary: Diagnosing Latent Inefficiency in Chain-of-Thought
Daeyeop Lee, Hwanjo Yu
The paper identifies and diagnoses inefficient reasoning steps in chain-of-thought prompting for large language models, introducing a benchmark and a training-free metric (CAID) to…
Automatic Open Knowledge Acquisition via Long Short-Term Memory Networks with Feedback Negative Sampling
Byungsoo Kim, Hwanjo Yu, Gary Geunbae Lee
Previous studies in Open Information Extraction (Open IE) are mainly based on extraction patterns. They manually define patterns or automatically learn them from a large corpus. Ho…
Global PIQA: Evaluating Commonsense Reasoning Across 100+ Languages and Cultures
Tyler A. Chang, Catherine Arnett, Abdelrahman Sadallah +377
To date, there exist almost no culturally-specific evaluation benchmarks for large language models (LLMs) that cover a large number of languages and cultures. In this paper, we pre…
Bidirectional Distillation for Top-K Recommender System
Wonbin Kweon, SeongKu Kang, Hwanjo Yu
Recommender systems (RS) have started to employ knowledge distillation, which is a model compression technique training a compact model (student) with the knowledge transferred fro…
STEPER: Step-wise Knowledge Distillation for Enhancing Reasoning Ability in Multi-Step Retrieval-Augmented Language Models
Kyumin Lee, Minjin Jeon, Sanghwan Jang +1
Answering complex real-world questions requires step-by-step retrieval and integration of relevant information to generate well-grounded responses. However, existing knowledge dist…
Unsupervised Proxy Selection for Session-based Recommender Systems
Junsu Cho, SeongKu Kang, Dongmin Hyun +1
Session-based Recommender Systems (SRSs) have been actively developed to recommend the next item of an anonymous short item sequence (i.e., session). Unlike sequence-aware recommen…
Unsupervised Differentiable Multi-aspect Network Embedding
Chanyoung Park, Carl Yang, Qi Zhu +3
Network embedding is an influential graph mining technique for representing nodes in a graph as distributed vectors. However, the majority of network embedding methods focus on lea…
Topology Distillation for Recommender System
SeongKu Kang, Junyoung Hwang, Wonbin Kweon +1
Recommender Systems (RS) have employed knowledge distillation which is a model compression technique training a compact student model with the knowledge transferred from a pre-trai…
PairSem: LLM-Guided Pairwise Semantic Matching for Scientific Document Retrieval
Wonbin Kweon, Runchu Tian, SeongKu Kang +4
Scientific document retrieval is a critical task for enabling knowledge discovery and supporting research across diverse domains. However, existing dense retrieval methods often st…
Dynamic Multi-Behavior Sequence Modeling for Next Item Recommendation
Junsu Cho, Dongmin Hyun, Dong won Lim +3
Sequential Recommender Systems (SRSs) aim to predict the next item that users will consume, by modeling the user interests within their item sequences. While most existing SRSs foc…
On the Effectiveness of Integration Methods for Multimodal Dialogue Response Retrieval
Seongbo Jang, Seonghyeon Lee, Dongha Lee +1
Multimodal chatbots have become one of the major topics for dialogue systems in both research community and industry. Recently, researchers have shed light on the multimodality of…
Collaborative Translational Metric Learning
Chanyoung Park, Donghyun Kim, Xing Xie +1
Recently, matrix factorization-based recommendation methods have been criticized for the problem raised by the triangle inequality violation. Although several metric learning-based…
How Diversely Can Language Models Solve Problems? Exploring the Algorithmic Diversity of Model-Generated Code
Seonghyeon Lee, Heejae Chon, Joonwon Jang +2
Language models (LMs) have exhibited impressive abilities in generating code from natural language requirements. In this work, we highlight the diversity of code generated by LMs a…
KoDialogBench: Evaluating Conversational Understanding of Language Models with Korean Dialogue Benchmark
Seongbo Jang, Seonghyeon Lee, Hwanjo Yu
As language models are often deployed as chatbot assistants, it becomes a virtue for models to engage in conversations in a user's first language. While these models are trained on…
Exploring Iterative Controllable Summarization with Large Language Models
Sangwon Ryu, Heejin Do, Daehee Kim +5
Large language models (LLMs) have demonstrated remarkable performance in abstractive summarization tasks. However, their ability to precisely control summary attributes (e.g., leng…
Beyond Learning from Next Item: Sequential Recommendation via Personalized Interest Sustainability
Dongmin Hyun, Chanyoung Park, Junsu Cho +1
Sequential recommender systems have shown effective suggestions by capturing users' interest drift. There have been two groups of existing sequential models: user- and item-centric…
Rectifying Demonstration Shortcut in In-Context Learning
Joonwon Jang, Sanghwan Jang, Wonbin Kweon +2
Large language models (LLMs) are able to solve various tasks with only a few demonstrations utilizing their in-context learning (ICL) abilities. However, LLMs often rely on their p…
Improving Scientific Document Retrieval with Academic Concept Index
Jeyun Lee, Junhyoung Lee, Wonbin Kweon +7
Adapting general-domain retrievers to scientific domains is challenging due to the scarcity of large-scale domain-specific relevance annotations and the substantial mismatch in voc…
From What to Respond to When to Respond: Timely Response Generation for Open-domain Dialogue Agents
Seongbo Jang, Minjin Jeon, Jaehoon Lee +3
While research on dialogue response generation has primarily focused on generating coherent responses conditioning on textual context, the critical question of when to respond grou…
DE-RRD: A Knowledge Distillation Framework for Recommender System
SeongKu Kang, Junyoung Hwang, Wonbin Kweon +1
Recent recommender systems have started to employ knowledge distillation, which is a model compression technique distilling knowledge from a cumbersome model (teacher) to a compact…
BHIN2vec: Balancing the Type of Relation in Heterogeneous Information Network
Seonghyeon Lee, Chanyoung Park, Hwanjo Yu
The goal of network embedding is to transform nodes in a network to a low-dimensional embedding vectors. Recently, heterogeneous network has shown to be effective in representing d…
Unsupervised Attributed Multiplex Network Embedding
Chanyoung Park, Donghyun Kim, Jiawei Han +1
Nodes in a multiplex network are connected by multiple types of relations. However, most existing network embedding methods assume that only a single type of relation exists betwee…
Capturing User Interests from Data Streams for Continual Sequential Recommendation
Gyuseok Lee, Hyunsik Yoo, Junyoung Hwang +2
Transformer-based sequential recommendation (SR) models excel at modeling long-range dependencies in user behavior via self-attention. However, updating them with continuously arri…
Federated Tensor Factorization for Computational Phenotyping
Yejin Kim, Jimeng Sun, Hwanjo Yu +1
Tensor factorization models offer an effective approach to convert massive electronic health records into meaningful clinical concepts (phenotypes) for data analysis. These models…
Out-of-Category Document Identification Using Target-Category Names as Weak Supervision
Dongha Lee, Dongmin Hyun, Jiawei Han +1
Identifying outlier documents, whose content is different from the majority of the documents in a corpus, has played an important role to manage a large text collection. However, d…
VLM2Rec: Resolving Modality Collapse in Vision-Language Model Embedders for Multimodal Sequential Recommendation
Junyoung Kim, Woojoo Kim, Jaehyung Lim +2
Sequential Recommendation (SR) in multimodal settings typically relies on small frozen pretrained encoders, which limits semantic capacity and prevents Collaborative Filtering (CF)…
REFIND at SemEval-2025 Task 3: Retrieval-Augmented Factuality Hallucination Detection in Large Language Models
DongGeon Lee, Hwanjo Yu
Hallucinations in large language model (LLM) outputs severely limit their reliability in knowledge-intensive tasks such as question answering. To address this challenge, we introdu…
Collaborative Diffusion Model for Recommender System
Gyuseok Lee, Yaochen Zhu, Hwanjo Yu +2
Diffusion-based recommender systems (DR) have gained increasing attention for their advanced generative and denoising capabilities. However, existing DR face two central limitation…
COMPASS: A Framework for Evaluating Organization-Specific Policy Alignment in LLMs
Dasol Choi, DongGeon Lee, Brigitta Jesica Kartono +6
As large language models are deployed in high-stakes enterprise applications, from healthcare to finance, ensuring adherence to organization-specific policies has become essential.…
Controlling Diversity at Inference: Guiding Diffusion Recommender Models with Targeted Category Preferences
Gwangseok Han, Wonbin Kweon, Minsoo Kim +1
Diversity control is an important task to alleviate bias amplification and filter bubble problems. The desired degree of diversity may fluctuate based on users' daily moods or busi…
Harmonic Dataset Distillation for Time Series Forecasting
Seungha Hong, Sanghwan Jang, Wonbin Kweon +3
Time Series forecasting (TSF) in the modern era faces significant computational and storage cost challenges due to the massive scale of real-world data. Dataset Distillation (DD),…
Delving into Instance-Dependent Label Noise in Graph Data: A Comprehensive Study and Benchmark
Suyeon Kim, SeongKu Kang, Dongwoo Kim +2
Graph Neural Networks (GNNs) have achieved state-of-the-art performance in node classification tasks but struggle with label noise in real-world data. Existing studies on graph lea…
Uncertainty Quantification and Decomposition for LLM-based Recommendation
Wonbin Kweon, Sanghwan Jang, SeongKu Kang +1
Despite the widespread adoption of large language models (LLMs) for recommendation, we demonstrate that LLMs often exhibit uncertainty in their recommendations. To ensure the trust…
Generating Multiple-Length Summaries via Reinforcement Learning for Unsupervised Sentence Summarization
Dongmin Hyun, Xiting Wang, Chanyoung Park +2
Sentence summarization shortens given texts while maintaining core contents of the texts. Unsupervised approaches have been studied to summarize texts without human-written summari…
Improving Scientific Document Retrieval with Concept Coverage-based Query Set Generation
SeongKu Kang, Bowen Jin, Wonbin Kweon +4
In specialized fields like the scientific domain, constructing large-scale human-annotated datasets poses a significant challenge due to the need for domain expertise. Recent metho…
Consensus Learning from Heterogeneous Objectives for One-Class Collaborative Filtering
SeongKu Kang, Dongha Lee, Wonbin Kweon +2
Over the past decades, for One-Class Collaborative Filtering (OCCF), many learning objectives have been researched based on a variety of underlying probabilistic models. From our a…
Click-aware purchase prediction with push at the top
Chanyoung Park, Donghyun Kim, Min-Chul Yang +2
Eliciting user preferences from purchase records for performing purchase prediction is challenging because negative feedback is not explicitly observed, and because treating all no…
Obtaining Calibrated Probabilities with Personalized Ranking Models
Wonbin Kweon, SeongKu Kang, Hwanjo Yu
For personalized ranking models, the well-calibrated probability of an item being preferred by a user has great practical value. While existing work shows promising results in imag…
Exploring Language Model's Code Generation Ability with Auxiliary Functions
Seonghyeon Lee, Sanghwan Jang, Seongbo Jang +2
Auxiliary function is a helpful component to improve language model's code generation ability. However, a systematic exploration of how they affect has yet to be done. In this work…
Multi-Domain Recommendation to Attract Users via Domain Preference Modeling
Hyunjun Ju, SeongKu Kang, Dongha Lee +3
Recently, web platforms have been operating various service domains simultaneously. Targeting a platform that operates multiple service domains, we introduce a new task, Multi-Doma…
Multi-Class Data Description for Out-of-distribution Detection
Dongha Lee, Sehun Yu, Hwanjo Yu
The capability of reliably detecting out-of-distribution samples is one of the key factors in deploying a good classifier, as the test distribution always does not match with the t…
Filling the Gaps: Selective Knowledge Augmentation for LLM Recommenders
Jaehyun Lee, Sanghwan Jang, SeongKu Kang +1
Large language models (LLMs) have recently emerged as powerful training-free recommenders. However, their knowledge of individual items is inevitably uneven due to imbalanced infor…
Continual Collaborative Distillation for Recommender System
Gyuseok Lee, SeongKu Kang, Wonbin Kweon +1
Knowledge distillation (KD) has emerged as a promising technique for addressing the computational challenges associated with deploying large-scale recommender systems. KD transfers…
MAP4TS: A Multi-Aspect Prompting Framework for Time-Series Forecasting with Large Language Models
Suchan Lee, Jihoon Choi, Sohyeon Lee +4
Recent advances have investigated the use of pretrained large language models (LLMs) for time-series forecasting by aligning numerical inputs with LLM embedding spaces. However, ex…
Improving Matrix Completion by Exploiting Rating Ordinality in Graph Neural Networks
Jaehyun Lee, SeongKu Kang, Hwanjo Yu
Matrix completion is an important area of research in recommender systems. Recent methods view a rating matrix as a user-item bi-partite graph with labeled edges denoting observed…
Bootstrapping User and Item Representations for One-Class Collaborative Filtering
Dongha Lee, SeongKu Kang, Hyunjun Ju +2
The goal of one-class collaborative filtering (OCCF) is to identify the user-item pairs that are positively-related but have not been interacted yet, where only a small portion of…
Q-Align: Alleviating Attention Leakage in Zero-Shot Appearance Transfer via Query-Query Alignment
Namu Kim, Wonbin Kweon, Minsoo Kim +1
We observe that zero-shot appearance transfer with large-scale image generation models faces a significant challenge: Attention Leakage. This challenge arises when the semantic map…
Learnable Structural Semantic Readout for Graph Classification
Dongha Lee, Su Kim, Seonghyeon Lee +2
With the great success of deep learning in various domains, graph neural networks (GNNs) also become a dominant approach to graph classification. By the help of a global readout op…
Task-Guided Pair Embedding in Heterogeneous Network
Chanyoung Park, Donghyun Kim, Qi Zhu +2
Many real-world tasks solved by heterogeneous network embedding methods can be cast as modeling the likelihood of pairwise relationship between two nodes. For example, the goal of…
BPL: Bias-adaptive Preference Distillation Learning for Recommender System
SeongKu Kang, Jianxun Lian, Dongha Lee +6
Recommender systems suffer from biases that cause the collected feedback to incompletely reveal user preference. While debiasing learning has been extensively studied, they mostly…
Top-Personalized-K Recommendation
Wonbin Kweon, SeongKu Kang, Sanghwan Jang +1
The conventional top-K recommendation, which presents the top-K items with the highest ranking scores, is a common practice for generating personalized ranking lists. However, is t…
Verbosity-Aware Rationale Reduction: Effective Reduction of Redundant Rationale via Principled Criteria
Joonwon Jang, Jaehee Kim, Wonbin Kweon +2
Large Language Models (LLMs) rely on generating extensive intermediate reasoning units (e.g., tokens, sentences) to enhance final answer quality across a wide range of complex task…
Weakly Supervised Temporal Anomaly Segmentation with Dynamic Time Warping
Dongha Lee, Sehun Yu, Hyunjun Ju +1
Most recent studies on detecting and localizing temporal anomalies have mainly employed deep neural networks to learn the normal patterns of temporal data in an unsupervised manner…
Eliciting Instruction-tuned Code Language Models' Capabilities to Utilize Auxiliary Function for Code Generation
Seonghyeon Lee, Suyeon Kim, Joonwon Jang +3
We study the code generation behavior of instruction-tuned models built on top of code pre-trained language models when they could access an auxiliary function to implement a funct…
Toward Interpretable Semantic Textual Similarity via Optimal Transport-based Contrastive Sentence Learning
Seonghyeon Lee, Dongha Lee, Seongbo Jang +1
Recently, finetuning a pretrained language model to capture the similarity between sentence embeddings has shown the state-of-the-art performance on the semantic textual similarity…
Learning Topology-Specific Experts for Molecular Property Prediction
Su Kim, Dongha Lee, SeongKu Kang +2
Recently, graph neural networks (GNNs) have been successfully applied to predicting molecular properties, which is one of the most classical cheminformatics tasks with various appl…
Personalized Federated Recommendation With Knowledge Guidance
Jaehyung Lim, Wonbin Kweon, Woojoo Kim +3
Federated Recommendation (FedRec) has emerged as a key paradigm for building privacy-preserving recommender systems. However, existing FedRec models face a critical dilemma: memory…
Taxonomy-guided Semantic Indexing for Academic Paper Search
SeongKu Kang, Yunyi Zhang, Pengcheng Jiang +3
Academic paper search is an essential task for efficient literature discovery and scientific advancement. While dense retrieval has advanced various ad-hoc searches, it often strug…
Learning Discriminative Dynamics with Label Corruption for Noisy Label Detection
Suyeon Kim, Dongha Lee, SeongKu Kang +3
Label noise, commonly found in real-world datasets, has a detrimental impact on a model's generalization. To effectively detect incorrectly labeled instances, previous works have m…
Deep Rating Elicitation for New Users in Collaborative Filtering
Wonbin Kweon, SeongKu Kang, Junyoung Hwang +1
Recent recommender systems started to use rating elicitation, which asks new users to rate a small seed itemset for inferring their preferences, to improve the quality of initial r…