Genetic Source Sensitivity and Transfer Learning in Genetic Programming

Helmuth, Thomas; Pantridge, Edward; Woolson, Grace; Spector, Lee

doi:10.1162/isal_a_00326

Cited by 15 publications

(19 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Research utilizing PSB1 in using transfer-learned instruction sets showed that the composition of the instruction set matters a great deal to problem-solving performance [16]. While we do not use fully transfer-learned instruction sets here, we do make use of one simple take-away: that including larger proportions of input instructions and constants/ERCs improves performance.…”

Section: Experimental Methods and System Parametersmentioning

confidence: 99%

Psb2

Helmuth

Kelly

2021

Proceedings of the Genetic and Evolutionary Computation Conference

Self Cite

View full text Add to dashboard Cite

For the past six years, researchers in genetic programming and other program synthesis disciplines have used the General Program Synthesis Benchmark Suite to benchmark many aspects of automatic program synthesis systems. These problems have been used to make notable progress toward the goal of general program synthesis: automatically creating the types of software that human programmers code. Many of the systems that have attempted the problems in the original benchmark suite have used it to demonstrate performance improvements granted through new techniques. Over time, the suite has gradually become outdated, hindering the accurate measurement of further improvements. The field needs a new set of more difficult benchmark problems to move beyond what was previously possible.In this paper, we describe the 25 new general program synthesis benchmark problems that make up PSB2, a new benchmark suite. These problems are curated from a variety of sources, including programming katas and college courses. We selected these problems to be more difficult than those in the original suite, and give results using PushGP showing this increase in difficulty. These new problems give plenty of room for improvement, pointing the way for the next six or more years of general program synthesis research. CCS CONCEPTS• Software and its engineering → Automatic programming.

show abstract

Section: Experimental Methods and System Parametersmentioning

confidence: 99%

Psb2

Helmuth

Kelly

2021

Proceedings of the Genetic and Evolutionary Computation Conference

Self Cite

View full text Add to dashboard Cite

show abstract

“…This means that a program can be supplemented or changed in several places without simply replacing large parts of the program. As the success rates could be notably improved with UMAD on several benchmark problems, it became the quasistandard mutation operator in PushGP for program synthesis (i.a., it is used in [40], [41], and [42]).…”

Section: A Stack-based Gpmentioning

confidence: 99%

“…In the next step, we replace <int> with min(<int>, <int>) as the relevant gene is 30 and the production rule (line 9) has four choices. The next gene (42) nests the second min() function inside the first one which leads to int5 = min(min(<int>, <int>), <int>) NEWLINE. With the next four genes (20, 0, 80, and 15), the nested min() function is filled with variables so the current state is int5 = min(min(int1, int2), <int>) NEWLINE.…”

Section: B Grammar-guided Gpmentioning

confidence: 99%

“…Small or Large [16], [29], [74], [48], [39], [75], [76], [40], [42], [41], [38] [11], [61], [79], [12], [2], [63] [68] 18…”

Section: Benchmark Problem Stack-based Gp Grammar-guided Gp Linear Gp...mentioning

confidence: 99%

“…For Loop Index [16], [29], [74], [48], [39], [75], [76], [40], [42], [38] [11], [61], [12], [2], [63] [68], [69], [70] 18…”

Section: Benchmark Problem Stack-based Gp Grammar-guided Gp Linear Gp...mentioning

confidence: 99%

See 2 more Smart Citations

Recent Developments in Program Synthesis with Evolutionary Algorithms

Sobania¹,

Schweim²,

Rothlauf³

2021

Preprint

View full text Add to dashboard Cite

The automatic generation of computer programs is one of the main applications with practical relevance in the field of evolutionary computation. With program synthesis techniques not only software developers could be supported in their everyday work but even users without any programming knowledge could be empowered to automate repetitive tasks and implement their own new functionality. In recent years, many novel program synthesis approaches based on evolutionary algorithms have been proposed and evaluated on common benchmark problems. Therefore, we identify in this work the relevant evolutionary program synthesis approaches and provide an in-depth analysis of their performance. The most influential approaches we identify are stack-based, grammar-guided, as well as linear genetic programming. Further, we find that these approaches perform well on benchmark problems if there is a simple mapping from the given input to the correct output. On problems where this mapping is complex, e.g., if the problem consists of several subproblems or requires iteration/recursion for a correct solution, results tend to be worse. Consequently, for future work, we encourage researchers not only to use a program's output for assessing the quality of a solution but also the way towards a solution (e.g., correctly solved sub-problems).

show abstract