Reusing Building Blocks of Extracted Knowledge to Solve Complex, Large-Scale Boolean Problems

@article{Iqbal2014ReusingBB,
  title={Reusing Building Blocks of Extracted Knowledge to Solve Complex, Large-Scale Boolean Problems},
  author={Muhammad Iqbal and Will N. Browne and Mengjie Zhang},
  journal={IEEE Transactions on Evolutionary Computation},
  year={2014},
  volume={18},
  pages={465-480}
}
Evolutionary computation techniques have had limited capabilities in solving large-scale problems due to the large search space demanding large memory and much longer training times. In the work presented here, a genetic programming like rich encoding scheme has been constructed to identify building blocks of knowledge in a learning classifier system. The fitter building blocks from the learning system trained against smaller problems have been utilized in a higher complexity problem in the… 
Reusing Learned Functionality to Address Complex Boolean Functions
TLDR
It is shown that through the reuse of learned functionality at the root and leaf nodes of code fragment trees, it is possible to solve complex problems such as the 18 bit hidden multiplexer problem.
Extending learning classifier system with cyclic graphs for scalability on complex, large-scale boolean problems
TLDR
This work introduces a state-machine based encoding scheme into scalable XCS, for the first time, in an attempt to develop a general scalable classifier system producing easily interpretable classifier rules.
A Layered Learning Approach to Scaling in Learning Classifier Systems for Boolean Problems
TLDR
Improvements to XCSCF* are proposed to enable it to be robust across multiple problem domains, and it is shown that from a conventional tabula rasa, it is possible to capture the general logic behind the tested domains, so the advanced system is capable of solving any individual n-bit Multiplexer, Carry-one, Majority-on, or Even-parity problem.
Extending XCS with Cyclic Graphs for Scalability on Complex Boolean Problems
TLDR
XCSSMA evolved, for the first time, compact and human readable general classifiers for the even-parity and carry problem domains, demonstrating its ability to produce scalable solutions using a cyclic representation.
Compaction for Code Fragment Based Learning Classifier Systems
TLDR
The new technique has been shown to create a new layer of DRs for the 70 Mux, something that the previous version was unable to accomplish in a timely manner, and it has been able to scale to more difficult problems in the same or a related domain.
Compaction for code fragment based learning classifier systems — Redux
TLDR
The new technique has been shown to create a new layer of DRs for the 70 Mux, something that the previous version was unable to accomplish in a timely manner, and it has been able to scale to more difficult problems in the same or a related domain.
Human-inspired Scaling in Learning Classifier Systems: Case Study on the n-bit Multiplexer Problem Set
TLDR
Results show that from a conventional tabula rasa, it is possible to learn a general solution to any n-bit Mux problem for the first time, and the developed approach is applicable to other scalable domains.
Integration of code-fragment based learning classifier systems for multiple domain perception and learning
TLDR
By utilizing the new CF system, it is possible to identify an unknown problem and to arrive at a viable solution and furthers the body of knowledge by tackling unknown problems (to the agent).
Common subtrees in related problems: A novel transfer learning approach for genetic programming
TLDR
The proposed non-random method for automatically finding relevant information for transfer between two source domain problems from the same problem domain based on common subtrees demonstrates ability as a general transfer learning technique for Genetic Programming.
A sensor tagging approach for reusing building blocks of knowledge in learning classifier systems
TLDR
A novel representation of classifier conditions for the Extended Classifier System (XCS) with the ST as the input representation is proposed, named Sensory Tag (ST), and the results indicate that the XCSSTC is capable of reusing building blocks of knowledge in the MUX problems.
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 80 REFERENCES
Extracting and using building blocks of knowledge in learning classifier systems
TLDR
The novel work presented here is to extract useful building blocks from smaller problems and reuse them to learn complex problems in the domain to improve effectiveness and reduces the number of training instances required in large problems, but requires more time due to more involved methods.
Extending learning classifier system with cyclic graphs for scalability on complex, large-scale boolean problems
TLDR
This work introduces a state-machine based encoding scheme into scalable XCS, for the first time, in an attempt to develop a general scalable classifier system producing easily interpretable classifier rules.
Evolving optimum populations with XCS classifier systems
TLDR
This paper identifies that consistency of action in rich alphabets guides LCS to optimum rule sets, and produces compact rule sets that can be easily converted to the optimum population.
Layered Learning in Boolean GP Problems
TLDR
This work has identified two approaches which, unlike previous methods, do not require prior understanding of a problem's functional decomposition into sub-goals and indicates that although one of the two approaches offers little advantage, the other leads to solution-finding performance significantly surpassing that of both conventional GP systems and those which incorporate automatically defined functions.
Investigating Scaling of an Abstracted LCS Utilising Ternary and S-Expression Alphabets
TLDR
This work shows that selection of appropriate S-Expression functions through domain knowledge improves scaling in problems, as expected, and is only a small step towards introducing abstraction to LCS.
Automatically defined functions for learning classifier systems
TLDR
This initial investigation of ADFs in LCS shows that the multiple genotypes to a phenotype issue in feature rich encodings disables the subsumption deletion function, which leads to much longer training times.
Problem Decomposition in Cartesian Genetic Programming
TLDR
In this chapter, a new approach called Embedded CGP (ECGP), is described that is capable of dynamically acquiring, evolving, and reusing modules to exploit modularity.
Automated Global Structure Extraction for Effective Local Building Block Processing in XCS
TLDR
The first competent LCSs, X CS/ECGA and XCS/BOA, that detect dependency structures online and propagate corresponding lower-level dependency structures effectively without any information about these structures given in advance are created.
Messy Genetic Algorithms: Motivation, Analysis, and First Results
TLDR
The mGA presented herein repeatedly achieves globally optimal results without prior knowledge of good string arrangements, and it does so at the very first generation in which strings are long enough to cover the problem.
Learning classifier systems: a complete introduction, review, and roadmap
TLDR
This paper aims to provide an accessible foundation for researchers of different backgrounds interested in selecting or developing their own LCS, including a simple yet thorough introduction, a historical review, and a roadmap of algorithmic components, emphasizing differences in alternative LCS implementations.
...
1
2
3
4
5
...