Dimitris Tsipras

Matrix Scaling and Balancing via Box Constrained Newton's Method and Interior Point Methods

In this paper, we study matrix scaling and balancing, which are fundamental problems in scientific computing, with a long line of work on them that dates back to the 1960s. We provide algorithms for both these problems that, ignoring logarithmic factors involving the dimension of the input matrix and the size of its entries, both run in time \widetilde{O}(m\log \kappa \log^2 (1/≥ilon)) where ≥ilon is the amount of error we are willing to tolerate. Here, \kappa represents the ratio between the largest and the smallest entries of the optimal scalings. This implies that our algorithms run in nearly-linear time whenever \kappa is quasi-polynomial, which includes, in particular, the case of strictly positive matrices. We complement our results by providing a separate algorithm that uses an interior-point method and runs in time \widetilde{O}(m^{3/2} \log (1/≥ilon)).In order to establish these results, we develop a new second-order optimization framework that enables us to treat both problems in a unified and principled manner. This framework identifies a certain generalization of linear system solving that we can use to efficiently minimize a broad class of functions, which we call second-order robust. We then show that in the context of the specific functions capturing matrix scaling and balancing, we can leverage and generalize the work on Laplacian system solving to make the algorithms obtained via this framework very efficient.

Efficient Money Burning in General Domains

We study mechanism design where the objective is to maximize the residual surplus, i.e., the total value of the outcome minus the payments charged to the agents, by truthful mechanisms. The motivation comes from applications where the payments charged are not in the form of actual monetary transfers, but take the form of wasted resources. We consider a general mechanism design setting with m discrete outcomes and n multidimensional agents. We present two randomized truthful mechanisms that extract an O(logm) fraction of the maximum social surplus as residual surplus. The first mechanism achieves an O(logm)-approximation to the social surplus, which is improved to an O(1)-approximation by the second mechanism. An interesting feature of the second mechanism is that it optimizes over an appropriately restricted space of probability distributions, thus achieving an efficient tradeoff between social surplus and the total amount of payments charged to the agents.