Researchers have introduced "double-scoring," a novel method for reliably extracting strong lottery tickets from large neural networks. This technique enhances the edge-popup approach by optimizing over an enlarged score space, which preserves access to original-coordinate masks. Experiments demonstrate that double-scoring significantly outperforms existing methods in strong-ticket extraction and shows reduced sensitivity to sparsity hyperparameters. AI
IMPACT Improves the efficiency and reliability of extracting sparse subnetworks from large neural networks, potentially leading to more efficient model training and deployment.
RANK_REASON The cluster contains a research paper detailing a new method for extracting lottery tickets from neural networks. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- arXiv
- Bryce Christopherson
- CatalyzeX
- DagsHub
- Double-Scoring
- edge-popup
- Hugging Face
- IArxiv
- lottery ticket hypothesis
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
