Learning classifier system with deep autoencoder

Matsumoto, K; Tajima, Yusuke; Saito, Rei; Nakata, Masaya; Satō, Hiroyuki; Kovacs, Tim; Takadama, Keiki

doi:10.1109/cec.2016.7744396

Cited by 5 publications

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, it is still limited in scalability as generalization is only provided through the "don't care" (#) symbols [9], which remove redundant/irrelevant features. In order to better partition the search space, i.e., not just in hyper-rectangles, alternative encodings have been proposed, e.g., neural networks [33], [34], [35], [36], [37], [38], tree-based programs [9], [39], [40], [41], [42], [43], and cyclic graphs [44].…”

Section: A Learning Classifier Systemsmentioning

confidence: 99%

ConCS: A Continual Classifier System for Continual Learning of Multiple Boolean Problems

Nguyen

Browne

Zhang

2023

IEEE Trans. Evol. Computat.

View full text Add to dashboard Cite

Human intelligence can simultaneously process many tasks with the ability to accumulate and reuse knowledge. Recent advances in artificial intelligence, such as transfer, multitask, and layered learning, seek to replicate these abilities. However, humans must specify the task order, which is often difficult particularly with uncertain domain knowledge. This work introduces a continual-learning system (ConCS), such that given an open-ended set of problems once each is solved its solution can contribute to solving further problems. The hypothesis is that the evolutionary computation approach of learning classifier systems (LCSs) can form this system due to its niched, cooperative rules. A collaboration of parallel LCSs identifies sets of patterns linking features to classes that can be reused in related problems automatically. Results from distinct Boolean and integer classification problems, with varying interrelations, show that by combining knowledge from simple problems, complex problems can be solved at increasing scales. 100% accuracy is achieved for the problems tested regardless of the order of task presentation. This includes intractable problems for previous approaches, e.g., n-bit Majority-on. A major contribution is that human guidance is now unnecessary to determine the task learning order. Furthermore, the system automatically generates the curricula for learning the most difficult tasks.

show abstract