18-23 July, 2022, Padova, Italy
Semantic Genetic Programming is a relatively new but rapidly growing research track of Genetic Programming (GP). As an evolutionary computation method, GP performs search and optimisation analogising natural evolution to produce programs to achieve the desired state. Program semantics, which provides grounding for the syntax of a program, is an important concept in computer programming. However, most GP methods manipulate programs only with syntax in mind and ignore the knowledge of their semantics. Recent works have shown that semantic GP which makes use of the semantics of GP programs induces a more aware version of GP. Semantics plays a crucial role in driving the evolutionary search, and thus semantic GP ends up being a more informed and intelligent method. With semantic awareness, GP dynamics are easier to understand and interpret, and inappropriate behaviours are easier to prevent or correct. This exactly contributes to a very important aspect of Explainable AI, which is a current hot research topic in the machine learning community.
Among these semantic GP methods, a relatively new variant is geometric semantic GP (GSGP) which presents a formal geometric framework for program semantics in different problem domains. The geometric framework also provides the basis for designing provably good semantic genetic operators for GP. GSGP aims to search the semantic space directly. The novel geometric semantic crossover and mutation operators act on the syntax of GP programs to produce offspring with desired semantic properties. The most obvious advantage of GSGP is on producing offspring that will not be worse than their worst parent. But more importantly, it induces a unimodal fitness landscape that facilitates the evolvability of GP. The unimodal fitness landscape with no local optimal works for any problems domains, which is a rare benefit in machine learning methods. Some deep investigations on the usefulness and properties of GSGP have demonstrated its advantage in complex real-world applications. A comprehensive review and introduction of these works will help to stimulate more research interests in this promising research track.
The tutorial structure:
Dr Qi Chen School of Engineering and Computer Science, Victoria University of Wellington, PO Box 600, Wellington, New Zealand. Qi.Chen@ecs.vuw.ac.nz Phone: +64-4-463 5233 x 8874; Fax: +64-4-463 5045.
Prof Bing Xue School of Engineering and Computer Science, Victoria University of Wellington, PO Box 600, Wellington, New Zealand. Bing.Xue@ecs.vuw.ac.nz Phone: +64-4-463 5542; Fax: +64-4-463 5045.
Prof Mengjie Zhang School of Engineering and Computer Science, Victoria University of Wellington, PO Box 600, Wellington, New Zealand. Mengjie.Zhang@ecs.vuw.ac.nz Phone: +64-4-463 5654; Fax: +64-4-463 5045
Dr Qi Chen is currently a Lecturer in School of Engineering and Computer Science at Victoria University of Wellington. Her research focuses mainly on data mining, machine learning, evolutionary computation, symbolic regression, feature manipulation. She has over 40 papers published in fully referred international journals and conferences and most of them are on symbolic modeling. Dr Chen has been serving as a program committee member of over ten international conferences including AAAI, IJCAI, IEEE CEC, IEEE SSCI, Australian AI and SEAL. She is serving as a reviewer of over ten international journals including IEEE Transactions on Cybernetics and IEEE Transactions on Evolutionary Computation.
Prof Bing Xue is currently a Professor in Artificial Intelligence and Deputy Head of School in the School of Engineering and Computer Science at Victoria University of Wellington. She has over 300 papers published in fully refereed international journals and conferences. She is currently the chair of IEEE CIS Evolutionary Computation Technical Committee, and Chair of IEEE CIS Task Force on Evolutionary Deep Learning and Applications, Vice-Chair (and founding chair) of IEEE CIS Task Force on Evolutionary Feature Selection and Construction, of IEEE Task Force on Transfer Learning & Transfer Optimization. She is an Associate Editor of several international journals, including IEEE TEVC. Prof Xue organised many special sessions and symposiums in international conferences such as IEEE WCCI/CEC, IEEE SSCI, and ACM GECCO. She has been a chair for many international conferences including program chair for SoCPaR2015 and Australasian AI 2018, finance chair for IEEE CEC 2019, general co-chair for IVCNZ 2020, workshop co-chair for IEEE ICDM 2021, and tutorial co-chair for WCCI 2022.
Prof Mengjie Zhang is a Fellow of Royal Society of New Zealand, a Fellow of Engineering New Zealand, a Fellow of IEEE, and currently Professor of Computer Science at Victoria University of Wellington, where he heads the interdisciplinary Evolutionary Computation Research Group. He has been serving as an associated editor or editorial board member for over 10 international journals including IEEE Transactions on Evolutionary Computation, and IEEE Transactions on Cybernetics. He is the Tutorial Chair for GECCO 2014, an AIS-BIO Track Chair for GECCO 2016, an EML Track Chair for GECCO 2017, and a GP Track Chair for GECCO 2020. Since 2012, he has been co-chairing several parts of IEEE CEC, SSCI, and EvoIASP/EvoApplications conference (he has been involving major EC conferences such as GECCO, CEC, EvoStar, SEAL). He has been co-organising and co-chairing many special sessions, and also delivered a keynote/plenary talk for IEEE CEC 2018, IEEE ICAVSS 2018, DOCSA 2019, IES 2017 and Chinese National Conference on AI in Law 2017.