Poster · Atlanta

SparDA: Sparse Decoupled Attention for Efficient Long-Context LLM Inference

Name: SparDA: Sparse Decoupled Attention for Efficient Long-Context LLM Inference
Start: 2026-12-09T16:30:00-05:00
End: 2026-12-09T19:30:00-05:00
Location: Hall C1, Georgia World Congress Center

When: Wed 9 Dec 2026, 16:30–19:30 EST
Where: Hall C1, Georgia World Congress Center
Session: Atlanta Poster Session 2
Authors: Yaosheng Fu, Guangxuan Xiao, Xin Dong, Song Han, Oreste Villa
Organisations: NVIDIA, Massachusetts Institute of Technology, Harvard University
QRS classification: Attention and transformers, Language models

Open in schedule Official page