Bumsub Ham* — accepted AI-conference papers

2024

FYI: Flip Your Images for Dataset Distillation

Byunggwan Son*, Youngmin Oh, Donghyeon Baek, Bumsub Ham*

ECCV 2024poster

"Dataset distillation synthesizes a small set of images from a large-scale real dataset such that synthetic and real images share similar behavioral properties (e.g, distributions of gradients or features) during a training process. Through extensive analyses on current methods and real datasets, to…

Cited by 3 Source ↗PDF ↗code ↗Scholar ↗

2024

Toward INT4 Fixed-Point Training via Exploring Quantization Error for Gradients

Dohyung Kim, Junghyup Lee, Jeimin Jeon, JAEHYEON MOON, Bumsub Ham*

ECCV 2024poster

"Network quantization generally converts full-precision weights and/or activations into low-bit fixed-point values in order to accelerate an inference process. Recent approaches to network quantization further discretize the gradients into low-bit fixed-point values, enabling an efficient training.…

Cited by 0 Source ↗PDF ↗Scholar ↗