Testable Learning of General Halfspaces under Massart Noise
This addresses a fundamental challenge in robust machine learning for scenarios with noisy data, providing a testable algorithm with theoretical guarantees, though it builds on prior work in testable learning.
The paper tackles the problem of testably learning general halfspaces under Massart noise with Gaussian marginals, achieving a quasi-polynomial time algorithm that matches known lower bounds for non-testable settings.
We study the algorithmic task of testably learning general Massart halfspaces under the Gaussian distribution. In the testable learning setting, the aim is the design of a tester-learner pair satisfying the following properties: (1) if the tester accepts, the learner outputs a hypothesis and a certificate that it achieves near-optimal error, and (2) it is highly unlikely that the tester rejects if the data satisfies the underlying assumptions. Our main result is the first testable learning algorithm for general halfspaces with Massart noise and Gaussian marginals. The complexity of our algorithm is $d^{\mathrm{polylog}(\min\{1/γ, 1/ε\})}$, where $ε$ is the excess error and $γ$ is the bias of the target halfspace, which qualitatively matches the known quasi-polynomial Statistical Query lower bound for the non-testable setting. The analysis of our algorithm hinges on a novel sandwiching polynomial approximation to the sign function with multiplicative error that may be of broader interest.