GRACE dynamically constructs and updates coresets for LLM training using representation diversity, gradient-based importance, and k-NN graph propagation to improve efficiency and performance.
Multi-Analyst Differential Privacy for Online Query Answering
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
verdicts
UNVERDICTED 2roles
background 1polarities
background 1representative citing papers
Big Bird enforces global device-epoch individual differential privacy for multi-querier Attribution by tying privacy-loss quotas to a stock-and-flow model of impressions and conversions with per-user-action caps.
citing papers explorer
-
GRACE: A Dynamic Coreset Selection Framework for Large Language Model Optimization
GRACE dynamically constructs and updates coresets for LLM training using representation diversity, gradient-based importance, and k-NN graph propagation to improve efficiency and performance.
-
Big Bird: Resilient Privacy Budgeting Across Untrusted Web Domains
Big Bird enforces global device-epoch individual differential privacy for multi-querier Attribution by tying privacy-loss quotas to a stock-and-flow model of impressions and conversions with per-user-action caps.