Replicable Learning of Large-Margin Halfspaces
This work provides incremental improvements in replicable learning algorithms for large-margin halfspaces, benefiting researchers in machine learning theory by enhancing efficiency and sample complexity.
The paper tackles the problem of learning large-margin halfspaces with replicable algorithms, achieving dimension-independent polynomial-time methods with optimal sample complexity for accuracy and improved sample complexity over prior work, including an SGD-based variant and an exponential-time algorithm with better sample complexity for margin parameters.
We provide efficient replicable algorithms for the problem of learning large-margin halfspaces. Our results improve upon the algorithms provided by Impagliazzo, Lei, Pitassi, and Sorrell [STOC, 2022]. We design the first dimension-independent replicable algorithms for this task which runs in polynomial time, is proper, and has strictly improved sample complexity compared to the one achieved by Impagliazzo et al. [2022] with respect to all the relevant parameters. Moreover, our first algorithm has sample complexity that is optimal with respect to the accuracy parameter $ε$. We also design an SGD-based replicable algorithm that, in some parameters' regimes, achieves better sample and time complexity than our first algorithm. Departing from the requirement of polynomial time algorithms, using the DP-to-Replicability reduction of Bun, Gaboardi, Hopkins, Impagliazzo, Lei, Pitassi, Sorrell, and Sivakumar [STOC, 2023], we show how to obtain a replicable algorithm for large-margin halfspaces with improved sample complexity with respect to the margin parameter $τ$, but running time doubly exponential in $1/τ^2$ and worse sample complexity dependence on $ε$ than one of our previous algorithms. We then design an improved algorithm with better sample complexity than all three of our previous algorithms and running time exponential in $1/τ^{2}$.