Solving Empirical Risk Minimization in the Current Matrix Multiplication Time
This provides a fast, deterministic solution for a broad class of machine learning optimization problems, generalizing recent linear programming results to more complex convex settings.
The paper tackles the problem of solving convex empirical risk minimization problems efficiently by developing an algorithm that runs in time matching the current matrix multiplication exponent, specifically O*(n^ω log(n/δ)), where ω ~ 2.38, with only logarithmic dependence on condition numbers.
Many convex problems in machine learning and computer science share the same form: \begin{align*} \min_{x} \sum_{i} f_i( A_i x + b_i), \end{align*} where $f_i$ are convex functions on $\mathbb{R}^{n_i}$ with constant $n_i$, $A_i \in \mathbb{R}^{n_i \times d}$, $b_i \in \mathbb{R}^{n_i}$ and $\sum_i n_i = n$. This problem generalizes linear programming and includes many problems in empirical risk minimization. In this paper, we give an algorithm that runs in time \begin{align*} O^* ( ( n^ω + n^{2.5 - α/2} + n^{2+ 1/6} ) \log (n / δ) ) \end{align*} where $ω$ is the exponent of matrix multiplication, $α$ is the dual exponent of matrix multiplication, and $δ$ is the relative accuracy. Note that the runtime has only a log dependence on the condition numbers or other data dependent parameters and these are captured in $δ$. For the current bound $ω\sim 2.38$ [Vassilevska Williams'12, Le Gall'14] and $α\sim 0.31$ [Le Gall, Urrutia'18], our runtime $O^* ( n^ω \log (n / δ))$ matches the current best for solving a dense least squares regression problem, a special case of the problem we consider. Very recently, [Alman'18] proved that all the current known techniques can not give a better $ω$ below $2.168$ which is larger than our $2+1/6$. Our result generalizes the very recent result of solving linear programs in the current matrix multiplication time [Cohen, Lee, Song'19] to a more broad class of problems. Our algorithm proposes two concepts which are different from [Cohen, Lee, Song'19] : $\bullet$ We give a robust deterministic central path method, whereas the previous one is a stochastic central path which updates weights by a random sparse vector. $\bullet$ We propose an efficient data-structure to maintain the central path of interior point methods even when the weights update vector is dense.