Knowledge reuse in genetic programming applied to visual learning

Abstract

We propose a method of knowledge reuse for an ensemble of genetic programming-based learners solving a visual learning task. First, we introduce a visual learning method that uses genetic programming individuals to represent hypotheses. Individuals-hypotheses process image representation composed of visual primitives derived from the training images that contain objects to be recognized. The process of recognition is generative, i.e., an individual is supposed to restore the shape of the processed object by drawing its reproduction on a separate canvas. This canonical method is extended with a knowledge reuse mechanism that allows a learner to import genetic material from hypotheses that evolved for the other decision classes (object classes). We compare the performance of the extended approach to the basic method on a real-world tasks of handwritten character recognition, and conclude that knowledge reuse leads to significant convergence speedup and, more importantly, significantly reduces the risk of overfitting.

Cross-task code reuse in genetic programming applied to visual learning

“…Here, we study a method that benefits from code fragments imported from other individuals. In comparison with our earlier work on this method (Jaśkowski et al, 2007b), here we additionally demonstrate its validity on a nontrivial, large multi-class problem, compare it to other machine learning techniques, analyze in-depth the effects of code reuse, provide extensive results for compound recognition systems, and place this study in the context of related research.…”

Multitask Visual Learning Using Genetic Programming

“…The idea of GP-based processing of attributed visual primitives was explored for the first time in Krawiec (2006), and was further developed in Jaśkowski (2006); Wieloch (2006); Jaśkowski et al (2007b); Krawiec (2007). In particular, in Jaśkowski et al (2007a) we examined the possibility of sequential knowledge sharing, as opposed to this contribution, where knowledge sharing takes place between learning processes that proceed in parallel.…”

Multi-task code reuse in genetic programming

“…In [9] we proposed a similar method of code reuse to this presented in this paper and evaluated it on a set of complicated pattern recognition problems. The method introduced in this paper has the advantage over the previous one of being parameterless.…”

“…This setup has the advantage of having only one parameter (the probability with which crossbreeding is engaged), however, other setups are also possible (see e.g. [9]). …”