Odjel za računalnu inteligenciju Hrvatske sekcije IEEE poziva Vas na predavanje:

Introduction to Genetic Programming and Geometric Semantic Genetic Programming

koje će održati dr. sc. Luca Manzoni (Assistant Professor at Dipartimento di Informatica, Sistemistica e Comunicazione, Università degli Studi di Milano-Bicocca) u utorak, 20. studenoga 2018. u 12:15 u prostoriji B5; predviđeno trajanje predavanja je 60 minuta.

Pozivamo i sve zainteresirane studente, a posebice studente kojima je zanimljiva problematika rješavanja optimizacijskih problema te pristupi evolucijskog računarstva da dođu na predavanje.

Više o predavanju dostupno je u opširnijem sadržaju obavijesti.

Sažetak: Genetic Programming (GP) is an evolutionary computation technique which is based on the idea of evolving computer programs in a way inspired by the principles of Darwinian evolution. Usually the quality of a solution is evaluated by looking at the semantics of the solution (i.e., the actual function computed by a program), while the evolutionary operators works on the syntax of the solution (i.e., the actual representation of the function).
In 2012 Moraglio et Al. introduced a new approach to GP: in Geometric Semantic GP (GSGP) the evolutionary operators have a specific effect on the semantics of the solutions that can be described in geometrical terms. In fields where geometric semantic operators are definable, GSGP proved effective; the use of subtree sharing also made it faster than traditional GP.
The talk will introduce GP and the concept of geometric and semantic operators. GSGP will be introduced with a discussion of both its advantages and disadvantages with respect to traditional GP.

 

O predavaču: Luca Manzoni received his PhD in computer science in 2013 from the University of Milano-Bicocca and he currently is an assistant professor at that university. In 2012 he received a short-term postdoctoral fellowship from JSPS and in 2017 he received an award from the University of Milan-Bicocca and the Accademia dei Lincei as the best postdoctoral researcher in computer science and mathematics at his university. He received multiple best paper awards and he is the authors of more than 70 research papers. His interests are related to bio-inspired models of computations, from cellular automata to membrane systems to genetic programming. While his main contributions are part of the more theoretical side of bio-inspired computing, he also collaborates to advance the application of these computational methods to real-life problems.

Autor: Marko Čupić
Popis obavijesti