Abstract
We give the first dimension-efficient algorithms for learning Rectified Linear Units (ReLUs), which are functions of the form $x \rightarrow \max(0, w \cdot x)$ with $w \in S^{n−1}$. Our algorithm works in the challenging Reliable Agnostic learning model of Kalai, Kanade, and Mansour [18] where the learner is given access to a distribution $\mathcal{D}$ on labeled examples but the labeling may be arbitrary. We construct a hypothesis that simultaneously minimizes the false-positive rate and the loss on inputs given positive labels by $\mathcal{D}$, for any convex, bounded, and Lipschitz loss function. The algorithm runs in polynomial-time (in n) with respect to any distribution on $S^{n−1}$ (the unit sphere in $n$ dimensions) and for any error parameter $\epsilon = \Omega(1/ \log n)$ (this yields a PTAS for a question raised by F. Bach on the complexity of maximizing ReLUs). These results are in contrast to known efficient algorithms for reliably learning linear threshold functions, where $\epsilon$ must be $\Omega(1)$ and strong assumptions are required on the marginal distribution. We can compose our results to obtain the first set of efficient algorithms for learning constant-depth networks of ReLUs. Our techniques combine kernel methods and polynomial approximations with a “dual-loss” approach to convex programming. As a byproduct we obtain a number of applications including the first set of efficient algorithms for “convex piecewise-linear fitting” and the first efficient algorithms for noisy polynomial reconstruction of low-weight polynomials on the unit sphere.
