8 papers
An Agentic Framework for Proactive Robots with Procedural Assistance Memory
Dohyun Lee, Minho Park, Jeonghoon Park +5
Proactive robots should anticipate user needs and provide assistance before being explicitly asked. A straightforward approach is to infer human intentions from commonsense knowled…
SharpDraft: Accelerating Long-Context Speculative Decoding with Cardinality-Aware Query Scaling
Jeonghoon Park, Seongwoon Jo, Jongwon Lee +1
Long-form reasoning makes inference expensive, and speculative decoding mitigates this cost by verifying multiple draft tokens in parallel. Its speedup, however, can fade as contex…
Fair Generation without Unfair Distortions: Debiasing Text-to-Image Generation with Entanglement-Free Attention
Jeonghoon Park, Juyoung Lee, Chaeyeon Chung +3
Recent advancements in diffusion-based text-to-image (T2I) models have enabled the generation of high-quality and photorealistic images from text. However, they often exhibit socie…
Enhancing Intrinsic Features for Debiasing via Investigating Class-Discerning Common Attributes in Bias-Contrastive Pair
Jeonghoon Park, Chaeyeon Chung, Juyoung Lee +1
In the image classification task, deep neural networks frequently rely on bias attributes that are spuriously correlated with a target class in the presence of dataset bias, result…
Revisiting the Importance of Amplifying Bias for Debiasing
Jungsoo Lee, Jeonghoon Park, Daeyoung Kim +3
In image classification, "debiasing" aims to train a classifier to be less susceptible to dataset bias, the strong correlation between peripheral attributes of data samples and a t…
Natural Attribute-based Shift Detection
Jeonghoon Park, Jimin Hong, Radhika Dua +4
Despite the impressive performance of deep networks in vision, language, and healthcare, unpredictable behaviors on samples from the distribution different than the training distri…
Deep Edge-Aware Interactive Colorization against Color-Bleeding Effects
Eungyeup Kim, Sanghyeon Lee, Jeonghoon Park +3
Deep neural networks for automatic image colorization often suffer from the color-bleeding artifact, a problematic color spreading near the boundaries between adjacent objects. Suc…
Evaluation of Out-of-Distribution Detection Performance of Self-Supervised Learning in a Controllable Environment
Jeonghoon Park, Kyungmin Jo, Daehoon Gwak +3
We evaluate the out-of-distribution (OOD) detection performance of self-supervised learning (SSL) techniques with a new evaluation framework. Unlike the previous evaluation methods…