We consider multitask learning of visual concepts within genetic programming (GP) framework. The proposed method evolves a population of GP individuals, with each of them composed of several GP trees that process visual primitives derived from input images. The two main trees are delegated to solving two different visual tasks and are allowed to share knowledge with each other by calling the remaining GP trees (subfunctions) included in the same individual. The method is applied to the visual learning task of recognizing simple shapes, using generative approach based on visual primitives, introduced in [17]. We compare this approach to a reference method devoid of knowledge sharing, and conclude that in the worst case cross-task learning performs equally well, and in many cases it leads to significant performance improvements in one or both solved tasks.
@INPROCEEDINGS { jaskowski07crosstask,
AUTHOR = { Wojciech Ja\'skowski and Krzysztof Krawiec and Bartosz Wieloch },
TITLE = { Genetic Programming for Cross-Task Knowledge Sharing },
BOOKTITLE = { GECCO '07: Proceedings of the 9th annual conference on Genetic and evolutionary computation },
YEAR = { 2007 },
EDITOR = { Dirk Thierens et al. },
VOLUME = { 2 },
PAGES = { 1620--1627 },
ADDRESS = { London },
MONTH = { jul },
ORGANIZATION = { ACM SIGEVO },
PUBLISHER = { ACM Press },
NOTE = { Nominated to the Best Paper Award in GP track },
ABSTRACT = { We consider multitask learning of visual concepts within genetic programming (GP) framework. The proposed method evolves a population of GP individuals, with each of them composed of several GP trees that process visual primitives derived from input images. The two main trees are delegated to solving two different visual tasks and are allowed to share knowledge with each other by calling the remaining GP trees (subfunctions) included in the same individual. The method is applied to the visual learning task of recognizing simple shapes, using generative approach based on visual primitives, introduced in [17]. We compare this approach to a reference method devoid of knowledge sharing, and conclude that in the worst case cross-task learning performs equally well, and in many cases it leads to significant performance improvements in one or both solved tasks. },
COMMENT = { 2 },
FILE = { jaskowski07genetic.pdf:\\jaskowski07genetic.pdf:PDF },
URL = { http://www.cs.bham.ac.uk/~wbl/biblio/gecco2007/docs/p1620.pdf },
ISBN = { 978-1-59593-698-1 },
KEYWORDS = { evolutionary computation, genetic programming, knowledge sharing, multitask learning, representation, visual learning },
ORDER = { 0 },
TIMESTAMP = { 2007.07.08 },
}