1. The complete title of one (or more) paper(s) published in the open literature describing the work that the author claims describe a human-competitive result; * A deep genetic programming based methodology for art media classification robust to adversarial perturbations https://doi.org/10.1007/978-3-030-64556-4_6 * Automated Design of Accurate and Robust Image Classifiers with Brain Programming https://doi.org/10.1145/3449726.3463179 This last is listed to appear at the ECADA workshop of GECCO at the following web page: https://bonsai.auburn.edu/ecada/GECCO2021/index.html 2. The name, complete physical mailing address, e-mail address, and phone number of EACH author of EACH paper(s); Gustavo Olague, CICESE Research Center, EvoVisión Laboratory, Ensenada, Carretera Ensenada-Tijuana 3918, Zona Playitas, Ensenada B.C. 22860, Mexico E-mail: olague@cicese.mx Gerardo Ibarra-Vazquez, Sierra Leona 307, Villa Real, Ciudad Victoria, Tamaulipas, Mexico, Zip Code 87027. E-mail: gerardo.ibarra.v@gmail.com Mariana Chan-Ley, CICESE Research Center, EvoVisión Laboratory, Ensenada, Carretera Ensenada-Tijuana 3918, Zona Playitas, Ensenada B.C. 22860, Mexico E-mail: mchan@cicese.edu.mx Cesar Puente, Facultad de Ingeniería Dr. Manuel Nava 8, Zona Universitaria, San Luis Potosí, San Luis Potosí, Mexico, Zip Code 78290. E-mail: cesar.puente@uaslp.mx Carlos Soubervielle-Montalvo, Facultad de Ingeniería Dr. Manuel Nava 8, Zona Universitaria, San Luis Potosí, San Luis Potosí, Mexico, Zip Code 78290. E-mail: carlos.soubervielle@uaslp.mx 3. The name of the corresponding author (i.e., the author to whom notices will be sent concerning the competition); Gustavo Olague 4. The abstract of the paper(s); * A deep genetic programming based methodology for art media classification robust to adversarial perturbations. Abstract: Art Media Classification problem is a current research area that has attracted attention due to the complex extraction and analysis of features of high-value art pieces. The perception of the attributes can not be subjective, as humans sometimes follow a biased interpretation of artworks while ensuring automated observation's trustworthiness. Machine Learning has outperformed many areas through its learning process of artificial feature extraction from images instead of designing handcrafted feature detectors. However, a major concern related to its reliability has brought attention because, with small perturbations made intentionally in the input image (adversarial attack), its prediction can be completely changed. In this manner, we foresee two ways of approaching the situation: (1) solve the problem of adversarial attacks in current neural networks methodologies, or (2) propose a different approach that can challenge deep learning without the effects of adversarial attacks. The first one has not been solved yet, and adversarial attacks have become even more complex to defend. Therefore, this work presents a Deep Genetic Programming method, called Brain Programming, that competes with deep learning and studies the transferability of adversarial attacks using two artworks databases made by art experts. The results show that the Brain Programming method preserves its performance in comparison with AlexNet, making it robust to these perturbations and competing to the performance of Deep Learning. * Automated Design of Accurate and Robust Image Classifiers with Brain Programming Abstract: Foster the mechanical design of artificial vision requires a delicate balance between high-level analytical methods and the discovery through metaheuristics of near-optimal functions working towards complex visual problems. Evolutionary computation and swarm intelligence have developed strategies that automatically design meaningful deep convolutional neural network architectures to create better image classifiers. However, these architectures have not surpassed handcraft models working with outdated problems with datasets of icon images. Nowadays, recent concerns about deep convolutional neural networks to adversarial attacks in the form of modifications to the input image can manipulate their output to make them untrustworthy. Brain programming is a hyper-heuristic whose aim is to work at a higher level of abstraction to develop automatically artificial visual cortex algorithms for a problem domain like image classification. This work's primary goal is to employ brain programming to design an artificial visual cortex to produce accurate and robust image classifiers in two problems. We analyze the final models designed by brain programming with the assumption of fooling the system using two adversarial attacks. In both experiments, brain programming constructed artificial brain models capable of competing with handcrafted deep convolutional neural networks without any influence in the predictions when an adversarial attack is present. 5. A list containing one or more of the eight letters (A, B, C, D, E, F, G, or H) that correspond to the criteria (see above) that the author claims that the work satisfies; (D) The result is publishable in its own right as a new scientific result independent of the fact that the result was mechanically created. (G) The result solves a problem of indisputable difficulty in its field. 6. A statement stating why the result satisfies the criteria that the contestant claims (see examples of statements of human-competitiveness as a guide to aid in constructing this part of the submission); The problem of adversarial attacks remains unsolved, according to the scientific literature. Many attacks demand a general solution since it is too complex to defend current deep learning systems. Therefore, we divide the eight criteria into three parts: 1. Criteria A, B, C, E, and F imply that a solution is already present; hence, we dismissed them from our list of claims since there are no solutions to compare our methodology. 2. Criteria H implies that existing competitions attempt to find defense mechanisms in current deep learning systems. Any methodology (e.g., evolutionary computation) based on such deep learning systems inherits this problem. Contests include challenges adapted to state-of-the-art image classification problems (e.g., hundreds of classes). Our proposed GP-like method (brain programming) requires instance problems whose solutions are competitive against deep learning techniques, where the provided defense mechanisms show their worth. 3. Criteria D and G are the only ones we fulfilled because our research was published in two different communities (visual computing and evolutionary computation). Our results show how designed solutions to complex problems (art media categorization and face recognition) with high-score performance are unaffected by several adversarial attacks (e.g., fast gradient sign method, and facial accessories perturbations). Such attacks are known as white box and black box attacks, and we implement them using different state of the art architectures (AlexNet and ResNet) for comparison. Even further experimentation about these studies claimed in the papers confirmed our results. The value of this research is to show a new research avenue to the evolutionary computation community, where the proposed algorithm looks for complex computational structures using only GP as the search mechanism, the solutions being robust to adversarial attacks, instead of the typical neuroevolution techniques that inherit such ill problem. 7. A full citation of the paper (that is, author names; title, publication date; name of journal, conference, or book in which article appeared; name of editors, if applicable, of the journal or edited book; publisher name; publisher city; page numbers, if applicable); * Olague G., Ibarra-Vázquez G., Chan-Ley M., Puente C., Soubervielle-Montalvo C., Martinez A. (2020) A Deep Genetic Programming Based Methodology for Art Media Classification Robust to Adversarial Perturbations. In: Bebis G. et al. (eds) Advances in Visual Computing. ISVC 2020. Lecture Notes in Computer Science, vol 12509. Springer, Cham. https://doi.org/10.1007/978-3-030-64556-4_6 * Ibarra-Vázquez G., Olague G., Chan-Ley M., Puente C., Soubervielle-Montalvo C. (2021). Automated Design of Accurate and Robust Image Classifiers with Brain Programming. In Proceedings of the 2021 Genetic and Evolutionary Computation Conference Companion (pp. 1-9). https://doi.org/10.1145/3449726.3463179 8. A statement either that "any prize money, if any, is to be divided equally among the co-authors" OR a specific percentage breakdown as to how the prize money, if any, is to be divided among the co-authors; Any prize money, if any, is for Gerardo Ibarra-Vazquez, and all authors agree on this. 9. A statement stating why the authors expect that their entry would be the "best," Recently, most researchers have focused their works on deep learning in areas such as computer vision. The evolutionary computation community is not the exception, as they have centered on developing strategies to search for meaningful deep learning architectures despite their inability to compete with them in problems such as image classification. However, the internal issues from deep learning are a big concern in the research community, risking their security and inherited to evolutionary computation solutions that use such technology. For example, an adversarial attack intentionally creates small perturbations in the input image to mislead a model to predict wrongly. These attacks are a trending topic in the research community because it has unveiled the reality on which the trustworthiness of deep learning results in compromised solutions, especially when applying it in critical areas of human activities. Current defense mechanisms proposed in the literature only provide specific solutions without solving the general problem. Therefore, we formulate a question: should we still focus on providing a solution to deep learning architectures, or can we find an alternative method that does not suffer this vulnerability? The results presented for consideration in this competition provide significant advancement to the state-of-the-art in this field and offer an alternative to deep learning in image classification. The study that we made regarding adversarial attacks for image classification demonstrate through different statistical and performance tests that our brain programming method based on evolutionary computation techniques and neuroscience theory competes with the state-of-the-art deep learning techniques in machine learning tasks, and we obtain a solution to the problem that is immune to adversarial attacks. 10. An indication of the general type of genetic or evolutionary computation used, such as GA (genetic algorithms), GP (genetic programming), ES (evolution strategies), EP (evolutionary programming), LCS (learning classifier systems), GI (genetic improvement), GE (grammatical evolution), GEP (gene expression programming), DE (differential evolution), etc. GP (Genetic Programming) 11. The date of publication of each paper. If the date of publication is not on or before the deadline for submission, but instead, the paper has been unconditionally accepted for publication and is "in press" by the deadline for this competition, the entry must include a copy of the documentation establishing that the paper meets the "in press" requirement. 07 December 2020 - A deep genetic programming based methodology for art media classification robust to adversarial perturbations "in press" - Automated Design of Accurate and Robust Image Classifiers with Brain Programming