Dynamic Algorithm Portfolios

Traditional Meta-Learning requires long training times, and is often focused on optimizing performance quality, neglecting computational complexity. Algorithm Portfolios are more robust, but present similar limitations. We reformulate algorithm selection as a time allocation problem: all candidate algorithms are run in parallel, and their relative priorities are continually updated based on runtime information, with the aim of minimizing the time to reach a desired performance level. Each algorithm’s priority is set based on its current time to solution, estimated according to a parametric model that is trained and used while solving a sequence of problems, gradually increasing its impact on the priority attribution. The use of censored sampling allows to train the model efficiently.

[1]  Michail G. Lagoudakis,et al.  Algorithm Selection using Reinforcement Learning , 2000, ICML.

[2]  Jürgen Schmidhuber,et al.  A Neural Network Model for Inter-problem Adaptive Online Time Allocation , 2005, ICANN.

[3]  John H. Holland,et al.  Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , 1992 .

[4]  Jürgen Schmidhuber,et al.  Adaptive Online Time Allocation to Search Algorithms , 2004, ECML.

[5]  Ricardo Vilalta,et al.  A Perspective View and Survey of Meta-Learning , 2002, Artificial Intelligence Review.

[6]  David Maxwell Chickering,et al.  A Bayesian Approach to Tackling Hard Computational Problems (Preliminary Report) , 2001, Electron. Notes Discret. Math..

[7]  P. W. Jones,et al.  Bandit Problems, Sequential Allocation of Experiments , 1987 .

[8]  Yoav Shoham,et al.  Learning the Empirical Hardness of Optimization Problems: The Case of Combinatorial Auctions , 2002, CP.

[9]  Ivana Kruijff-Korbayová,et al.  A Portfolio Approach to Algorithm Selection , 2003, IJCAI.

[10]  Jürgen Schmidhuber,et al.  Optimal Ordered Problem Solver , 2002, Machine Learning.

[11]  Heekuck Oh,et al.  Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[12]  Hilan Bensusan,et al.  Meta-Learning by Landmarking Various Learning Algorithms , 2000, ICML.

[13]  Donald A. Berry,et al.  Bandit Problems: Sequential Allocation of Experiments. , 1986 .

[14]  Shlomo Zilberstein,et al.  Monitoring and control of anytime algorithms: A dynamic programming approach , 2001, Artif. Intell..

[15]  Fernando G. Lobo,et al.  A parameter-less genetic algorithm , 1999, GECCO.

[16]  Marek Petrik,et al.  Statistically Optimal Combination of Algorithms , 2004 .

[17]  Andrew W. Moore,et al.  Efficient Algorithms for Minimizing Cross Validation Error , 1994, ICML.

[18]  Andrew W. Moore,et al.  Reinforcement Learning: A Survey , 1996, J. Artif. Intell. Res..

[19]  R. Solomonoff.  Progress In Incremental Machine Learning , 2003 .

[20]  Wayne Nelson,et al.  Applied life data analysis , 1983 .

[21]  Jürgen Schmidhuber,et al.  Shifting Inductive Bias with Success-Story Algorithm, Adaptive Levin Search, and Incremental Self-Improvement , 1997, Machine Learning.

[22]  R. Geoff Dromey,et al.  An algorithm for the selection problem , 1986, Softw. Pract. Exp..

[23]  Bart Selman,et al.  Algorithm portfolios , 2001, Artif. Intell..

[24]  Mark S. Boddy,et al.  Deliberation Scheduling for Problem Solving in Time-Constrained Environments , 1994, Artif. Intell..