Alert button
Picture for Hwanjun Song

Hwanjun Song

Alert button

TofuEval: Evaluating Hallucinations of LLMs on Topic-Focused Dialogue Summarization

Feb 20, 2024
Liyan Tang, Igor Shalyminov, Amy Wing-mei Wong, Jon Burnsky, Jake W. Vincent, Yu'an Yang, Siffi Singh, Song Feng, Hwanjun Song, Hang Su, Lijia Sun, Yi Zhang, Saab Mansour, Kathleen McKeown

Viaarxiv icon

Adaptive Shortcut Debiasing for Online Continual Learning

Dec 14, 2023
Doyoung Kim, Dongmin Park, Yooju Shin, Jihwan Bang, Hwanjun Song, Jae-Gil Lee

Figure 1 for Adaptive Shortcut Debiasing for Online Continual Learning
Figure 2 for Adaptive Shortcut Debiasing for Online Continual Learning
Figure 3 for Adaptive Shortcut Debiasing for Online Continual Learning
Figure 4 for Adaptive Shortcut Debiasing for Online Continual Learning
Viaarxiv icon

Toward Robustness in Multi-label Classification: A Data Augmentation Strategy against Imbalance and Noise

Dec 12, 2023
Hwanjun Song, Minseok Kim, Jae-Gil Lee

Viaarxiv icon

One Size Fits All for Semantic Shifts: Adaptive Prompt Tuning for Continual Learning

Nov 18, 2023
Doyoung Kim, Susik Yoon, Dongmin Park, Youngjun Lee, Hwanjun Song, Jihwan Bang, Jae-Gil Lee

Figure 1 for One Size Fits All for Semantic Shifts: Adaptive Prompt Tuning for Continual Learning
Figure 2 for One Size Fits All for Semantic Shifts: Adaptive Prompt Tuning for Continual Learning
Figure 3 for One Size Fits All for Semantic Shifts: Adaptive Prompt Tuning for Continual Learning
Figure 4 for One Size Fits All for Semantic Shifts: Adaptive Prompt Tuning for Continual Learning
Viaarxiv icon

Robust Data Pruning under Label Noise via Maximizing Re-labeling Accuracy

Nov 02, 2023
Dongmin Park, Seola Choi, Doyoung Kim, Hwanjun Song, Jae-Gil Lee

Viaarxiv icon

Enhancing Abstractiveness of Summarization Models through Calibrated Distillation

Oct 20, 2023
Hwanjun Song, Igor Shalyminov, Hang Su, Siffi Singh, Kaisheng Yao, Saab Mansour

Figure 1 for Enhancing Abstractiveness of Summarization Models through Calibrated Distillation
Figure 2 for Enhancing Abstractiveness of Summarization Models through Calibrated Distillation
Figure 3 for Enhancing Abstractiveness of Summarization Models through Calibrated Distillation
Figure 4 for Enhancing Abstractiveness of Summarization Models through Calibrated Distillation
Viaarxiv icon

Fast and Robust Early-Exiting Framework for Autoregressive Language Models with Synchronized Parallel Decoding

Oct 09, 2023
Sangmin Bae, Jongwoo Ko, Hwanjun Song, Se-Young Yun

Viaarxiv icon

Prompt-Guided Transformers for End-to-End Open-Vocabulary Object Detection

Mar 25, 2023
Hwanjun Song, Jihwan Bang

Figure 1 for Prompt-Guided Transformers for End-to-End Open-Vocabulary Object Detection
Figure 2 for Prompt-Guided Transformers for End-to-End Open-Vocabulary Object Detection
Figure 3 for Prompt-Guided Transformers for End-to-End Open-Vocabulary Object Detection
Figure 4 for Prompt-Guided Transformers for End-to-End Open-Vocabulary Object Detection
Viaarxiv icon

Q-HyViT: Post-Training Quantization for Hybrid Vision Transformer with Bridge Block Reconstruction

Mar 22, 2023
Jemin Lee, Yongin Kwon, Jeman Park, Misun Yu, Hwanjun Song

Figure 1 for Q-HyViT: Post-Training Quantization for Hybrid Vision Transformer with Bridge Block Reconstruction
Figure 2 for Q-HyViT: Post-Training Quantization for Hybrid Vision Transformer with Bridge Block Reconstruction
Figure 3 for Q-HyViT: Post-Training Quantization for Hybrid Vision Transformer with Bridge Block Reconstruction
Figure 4 for Q-HyViT: Post-Training Quantization for Hybrid Vision Transformer with Bridge Block Reconstruction
Viaarxiv icon

Re-thinking Federated Active Learning based on Inter-class Diversity

Mar 22, 2023
SangMook Kim, Sangmin Bae, Hwanjun Song, Se-Young Yun

Figure 1 for Re-thinking Federated Active Learning based on Inter-class Diversity
Figure 2 for Re-thinking Federated Active Learning based on Inter-class Diversity
Figure 3 for Re-thinking Federated Active Learning based on Inter-class Diversity
Figure 4 for Re-thinking Federated Active Learning based on Inter-class Diversity
Viaarxiv icon