- Author:
- Elad Hazan,
- Adviser:
- Sanjeev Arora
In this thesis we study algorithms for online convex optimization and their relation to approximate optimization.
In the first part, we propose a new algorithm for a general online optimization framework called online convex optimization. Whereas previous efficient algorithms are mostly gradient-descent based, the new algorithm is inspired by the Newton-Raphson method for convex optimization, and hence called O NLINE N EWTON S TEP . We prove that in certain scenarios O NLINE N EWTON S TEP guarantees logarithmic regret, as opposed to polynomial bounds achieved by previous algorithms. The analysis is based on new insights concerning the natural "follow-the-leader" method for online optimization, answers some open problems regarding the latter.
One application is for the portfolio management problem, for which we describe experimental results over real market data.
In the second part of the thesis, we describe a general scheme of utilizing online game playing algorithms to obtain efficient algorithms for offline optimization. Using new and old online convex optimization algorithms we show how to derive the following: (1) Approximation algorithms for convex programming with linear dependence on the approximation guaranty. (2) Fast algorithms for approximate Semidefinite Programming. (3) Efficient algorithms for haplotype frequency estimation.
Cited By
- Elbassioni K, Makino K and Najy W (2019). A Multiplicative Weight Updates Algorithm for Packing and Covering Semi-infinite Linear Programs, Algorithmica, 81:6, (2377-2429), Online publication date: 1-Jun-2019.
Li B and Hoi S (2014). Online portfolio selection, ACM Computing Surveys, 46:3, (1-36), Online publication date: 1-Jan-2014.- Gofer E and Mansour Y Lower bounds on individual sequence regret Proceedings of the 23rd international conference on Algorithmic Learning Theory, (275-289)
- Veness J, Sunehag P and Hutter M On ensemble techniques for AIXI approximation Proceedings of the 5th international conference on Artificial General Intelligence, (341-351)
- Li B, Hoi S and Gopalkrishnan V (2011). CORN, ACM Transactions on Intelligent Systems and Technology (TIST), 2:3, (1-29), Online publication date: 1-Apr-2011.
- Hazan E, Agarwal A and Kale S (2007). Logarithmic regret algorithms for online convex optimization, Machine Language, 69:2-3, (169-192), Online publication date: 1-Dec-2007.
Index Terms
- Efficient algorithms for online convex optimization and their applications
Recommendations
An online convex optimization-based framework for convex bilevel optimization
AbstractWe propose a new framework for solving the convex bilevel optimization problem, where one optimizes a convex objective over the optimal solutions of another convex optimization problem. As a key step of our framework, we form an online convex ...
A powerful and efficient algorithm for numerical function optimization: artificial bee colony (ABC) algorithm
Swarm intelligence is a research branch that models the population of interacting agents or swarms that are able to self-organize. An ant colony, a flock of birds or an immune system is a typical example of a swarm system. Bees' swarming around their ...