Month: September 2022

BLOAT EFFECTS. "It has been long noticed that there are emergent phenomena in genetic programming (GP) runs unintended by the human designer of the algorithm. Early on it was observed that code which does not change the output of the program (i.e. non-effective code) appears in many GP runs. It was also noted that bloat affects many GP systems. Reasons for bloat and non-effective code have been examined in years past and remedies have been developed more or less effective under particular circumstances." https://lnkd.in/eABn2vx View in LinkedIn

REPURPOSED CODE. "In bloated programs the small number of effective or expressed instructions appear in both repeated and non-repeated code. Hinting that building-blocks or code reuse may evolve in unplanned ways. Non-effective code and bloat are only the tip of an iceberg and that there is more to be discovered about “emergent phenomena” in GP runs." https://lnkd.in/eABn2vx View in LinkedIn

“EMERGENT INTELLIGENCE, a method of problem solving that allows task specific knowledge to emerge from the interaction of the problem solver and a task description, here represented by the fitness function.” http://scholar.google.es/scholar_url?url=http://citeseerx.ist.psu.edu/viewdoc/download%3Fdoi%3D10.1.1.15.5594%26rep%3Drep1%26type%3Dpdf&hl=en&sa=X&scisig=AAGBfm0J-z_xL5_kjwEhLP0aVOrhbAtTfA&nossl=1&oi=scholarr&ved=0ahUKEwi3wpKxxs_MAhXG2B4KHXRODt0QgAMICigBMAA View in LinkedIn

PROBLEM SOLVING. "Problem solving in AI takes one of two forms: computational models of humans solving problems and the construction of programs that solve problems as well or better than humans. Clearly, the techniques of genetic programming and all evolutionary computations fit into the latter methodology of building intelligent programs." http://scholar.google.es/scholar_url?url=http://citeseerx.ist.psu.edu/viewdoc/download%3Fdoi%3D10.1.1.15.5594%26rep%3Drep1%26type%3Dpdf&hl=en&sa=X&scisig=AAGBfm0J-z_xL5_kjwEhLP0aVOrhbAtTfA&nossl=1&oi=scholarr&ved=0ahUKEwi3wpKxxs_MAhXG2B4KHXRODt0QgAMICigBMAA View in LinkedIn

STRONG AND WEAK METHODS. "AI problem solving methods divides into two distinct categories: weak and strong. A strong method is one that contains a significant amount of task specific knowledge in order to solve it. In contrast, a weak method requires little or no knowledge about a task. As the name implies, strong methods are more powerful algorithms than weak methods." http://scholar.google.es/scholar_url?url=http://citeseerx.ist.psu.edu/viewdoc/download%3Fdoi%3D10.1.1.15.5594%26rep%3Drep1%26type%3Dpdf&hl=en&sa=X&scisig=AAGBfm0J-z_xL5_kjwEhLP0aVOrhbAtTfA&nossl=1&oi=scholarr&ved=0ahUKEwi3wpKxxs_MAhXG2B4KHXRODt0QgAMICigBMAA View in LinkedIn

OPERATIONS. "All evolutionary computations perform a simple SELECT-COPY-MODIFY loop over the population each generation. SELECT identifies the number of copies each member of the current population produces in the next population based on a relative ranking by the fitness function. COPY makes identical copies of the specified numbers for each member. Finally, MODIFY applies the reproductive operators." http://scholar.google.es/scholar_url?url=http://citeseerx.ist.psu.edu/viewdoc/download%3Fdoi%3D10.1.1.15.5594%26rep%3Drep1%26type%3Dpdf&hl=en&sa=X&scisig=AAGBfm0J-z_xL5_kjwEhLP0aVOrhbAtTfA&nossl=1&oi=scholarr&ved=0ahUKEwi3wpKxxs_MAhXG2B4KHXRODt0QgAMICigBMAA View in LinkedIn

TABLAR RAZAR. "Emergent intelligence (EI) is a method of reducing or removing the reliance on explicit knowledge in a problem solver by favoring instead the emergence of problem specific constraints as a direct consequence of the action of problem solving." http://scholar.google.es/scholar_url?url=http://citeseerx.ist.psu.edu/viewdoc/download%3Fdoi%3D10.1.1.15.5594%26rep%3Drep1%26type%3Dpdf&hl=en&sa=X&scisig=AAGBfm0J-z_xL5_kjwEhLP0aVOrhbAtTfA&nossl=1&oi=scholarr&ved=0ahUKEwi3wpKxxs_MAhXG2B4KHXRODt0QgAMICigBMAA View in LinkedIn

CONTEXTUAL RESOURCES. "Using emergent intelligence allows the removal of explicit knowledge that is a natural consequence of the problem solving process interacting with the task environment. By allowing the task environment to be an integral component of the problem solving algorithm, all the natural constraints, including those too subtle for a knowledge engineer to extract, are available to the algorithm and emerge at appropriate moments while solving problems." http://scholar.google.es/scholar_url?url=http://citeseerx.ist.psu.edu/viewdoc/download%3Fdoi%3D10.1.1.15.5594%26rep%3Drep1%26type%3Dpdf&hl=en&sa=X&scisig=AAGBfm0J-z_xL5_kjwEhLP0aVOrhbAtTfA&nossl=1&oi=scholarr&ved=0ahUKEwi3wpKxxs_MAhXG2B4KHXRODt0QgAMICigBMAA View in LinkedIn

PROGRAMMING INTRONS. "Emergence of Introns in GP. These components are unnecessary since they can be removed from the program without altering the solution the program represents. The ubiquity of redundant constructs suggests a more basic truth about their nature. Introducing introns into a genetic algorithm’s representation improves the success rate as much as a factor of 10." (GP = genetic programming). http://scholar.google.es/scholar_url?url=http://citeseerx.ist.psu.edu/viewdoc/download%3Fdoi%3D10.1.1.15.5594%26rep%3Drep1%26type%3Dpdf&hl=en&sa=X&scisig=AAGBfm0J-z_xL5_kjwEhLP0aVOrhbAtTfA&nossl=1&oi=scholarr&ved=0ahUKEwi3wpKxxs_MAhXG2B4KHXRODt0QgAMICigBMAA View in LinkedIn

REPETITIONS. "There is an advantage from the viewpoint of the evolving genetic program for syntactically redundant constructions that are semantically transparent. Because the redundant functions do not effect the semantics of the sub-expression, if a crossover operation is performed such that it splits the redundant components, the semantics of the removed subtree are still preserved." https://lnkd.in/dcK2gx8 View in LinkedIn