Jaeyoung Do

2 papers indexed

arxivcs.DCcs.AIcs.ARcs.LG2026-07-24

Unified Static-Dynamic Pruning for Efficient LLM Inference

The increasing deployment of large language models (LLMs) has magnified the computational and memory bottlenecks of autoregressive decoding, where low compute intensity and bandwidth-bound kernels dominate inference cost. Weight pruning offers a promising remedy, but existing met…

View free PDFSource page

arxivcs.LG2026-07-01

SUM: Unified Geometric Surgery on Spatio-Temporal Adaptation Vectors for Federated Class Incremental Learning

Jaeik Kim, Jaeyoung Do

Real-world intelligent systems often require both distributed collaboration across data-isolated clients and continual adaptation to evolving tasks. This setting naturally gives rise to Federated Class Incremental Learning (FCIL), which combines Federated Learning (FL) and Contin…

View free PDFSource page