Private online learning and prediction for Littlestone classes
What it is
This paper investigates the mistake bounds for differentially private online learning and prediction under oblivious realizable adversaries. It establishes a separation between the sample complexity of private learning and private prediction, showing that learning requires a mistake bound growing with the time horizon while prediction admits a bound independent of it for classes of finite Littlestone dimension.
The problem it solves
The lack of tight lower bounds for private online learning in specific privacy regimes and the unclear distinction in sample complexity between learning and prediction tasks.
What you could use it for
- Theoretical analysis of differential privacy in online learning algorithms
- Designing private prediction systems that avoid releasing hypotheses
- Benchmarking the efficiency of private learning versus private prediction
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.