%B Lecture Notes in Computer Science
%I Springer
%P 27-41
%L discovery10094843
%K Search-Based Software Testing · Software Engineering · Fitness Function · Machine Learning · Neural Networks
%V 11664
%C Tallinn, Estonia
%J Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
%E S Nejati
%E G Gay
%A L Joffe
%A D Clark
%X A central requirement for any Search-Based Software Testing (SBST) technique is a convenient and meaningful fitness landscape. Whether one follows a targeted or a diversification driven strategy, a search landscape needs to be large, continuous, easy to construct and representative of the underlying property of interest. Constructing such a landscape is not a trivial task often requiring a significant manual effort by an expert.

We present an approach for constructing meaningful and convenient fitness landscapes using neural networks (NN) – for targeted and diversification strategies alike. We suggest that output of an NN predictor can be interpreted as a fitness for a targeted strategy. The NN is trained on a corpus of execution traces and various properties of interest, prior to searching. During search, the trained NN is queried to predict an estimate of a property given an execution trace. The outputs of the NN form a convenient search space which is strongly representative of a number of properties. We believe that such a search space can be readily used for driving a search towards specific properties of interest.

For a diversification strategy, we propose the use of an autoencoder; a mechanism for compacting data into an n-dimensional “latent” space. In it, datapoints are arranged according to the similarity of their salient features. We show that a latent space of execution traces possesses characteristics of a convenient search landscape: it is continuous, large and crucially, it defines a notion of similarity to arbitrary observations.
%O This version is the author accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions.
%D 2019
%T Constructing Search Spaces for Search-Based Software Testing Using Neural Networks