No description available.

Artificial intelligence has traditionally been approached through centralized architectures and optimization of specific metrics on large datasets. However, the frontiers of fields spanning cognitive science, biology, physics, and computer science suggest that intelligence is better understood as a multi-scale, decentralized, emergent phenomenon. As such, scaling up approaches that mirror the natural world may be one of the next big advances in AI. This thesis presents Coralai, a framework for efficiently simulating the emergence of diverse artificial life ecosystems integrated with modular physics. The key innovations of Coralai include: 1) Hosting diverse Neural Cellular Automata organisms in the same simulation that can interact and evolve; 2) Allowing user-defined physics and weather that organisms adapt to and can utilize to enact environmental changes; 3) Hardware-acceleration using Taichi, PyTorch, and HyperNEAT, enabling interactive evolution of ecosystems with 500k evolved parameters on a grid of 1m+ 16-channel physics-governed cells, all in real-time on a laptop. Initial experiments with Coralai demonstrate the emergence of diverse ecosystems of organisms that employ a variety of strategies to compete for resources in dynamic environments. Key observations include competing mobile and sessile organisms, organisms that exploit environmental niches like dense energy sources, and cyclic dynamics of greedy dominance out-competed by resilience.

Made available in DSpace on 2016-04-29T14:23:52Z (GMT). No. of bitstreams: 1 Davi Marcos Machado.pdf: 448497 bytes, checksum: 9f7a19fb13d50b4a4a08e568c575dd89 (MD5) Previous issue date: 2009-05-26 / Ending of century XX and mainly from the beginning of the XXI century, the popularization of the Internet access and the emergence of software which provide friendly environment for the inter-relationship of people in virtual space (MySpace, Orkut, Blogspot etc), has made the cyberspace to be occupied by a huge number of people with more diverse interests, which resulted in the appearance of virtual communities or social networks. These communities or networks are organized around several themes and interests. May simply be groups of adolescents who have fun to exchange views on some prominent personality in the media to communities that discuss issues such as education, philosophy, public health or groups that develop a project of great social or economic relevance. The objective of this research is to evaluate, through empirical analysis, based on the available literature, the feasibility of classifying the millions of virtual communities or social networks in just three major groups: virtual communities or social networks organized around projects, people and ideas. We believe that these three major groups are able to include all types of social groups in cyberspace, considering that each one has its peculiarities of organization and structure. "Project means a temporary business or a complex sequence of activities to provide a product. For people we considered the charismatic individual, able to attract other individuals to their environment, either by their capacity for leadership or the ability to provide interest information to a group of people. Ideas is meant a concept, a set of thoughts of an individual or a social group on any one topic, theory, point of view or opinion. The purpose of this classification (projects, people and ideas) is to identify the different forms of grouping, outlining the main features of each of these groups, the structure of the environment in which it is organized and dynamic. These settings can help people, institutions or companies who wish to use the cyberspace how environment of knowledge generation or development of activities or simply exchange ideas. From the characteristics of each type of community is possible to define the best structure and environment to be built, considering the interests and goals of those who structure the community or the network / No final do século XX e no início do século XXI, a popularização do acesso a Internet e o surgimento de softwares amigáveis e ambientes propícios à inter-relação de pessoas no espaço virtual (MySpace, Orkut, Blogspot etc) fez com que o ciberespaço passasse a ser povoado por um número imenso de pessoas com os mais diversos interesses, que resultou no surgimento de comunidades virtuais ou de redes sociais. Essas comunidades ou redes se organizam em torno dos mais diversos interesses. Podem ser grupos de adolescentes que se divertem ao trocar impressões sobre alguma personalidade em destaque na mídia até comunidades que debatem temas como educação, filosofia, saúde pública ou mesmo grupos que desenvolvem algum projeto de grande relevância social ou econômica. O objetivo desta pesquisa é avaliar a viabilidade de se classificar os milhões de comunidades virtuais ou redes sociais em apenas três grandes agrupamentos: comunidades ou redes organizadas em torno de projetos , de pessoas e de idéias . Acreditamos que esses três grandes grupos são capazes de comportar todos os tipos de agrupamentos sociais no ciberespaço, sendo que cada um deles possui suas peculiaridades de organização e estrutura. Entendemos por projeto um ação temporária que tempo por objetivo fornecer um produto singular. Por pessoas consideramos o individuo carismático, com capacidade de atração de outros indivíduos para seu entorno; e por idéias se entende um conceito, um conjunto de pensamentos ou concepções de um indivíduo ou de um grupo social sobre um tema qualquer. O objetivo de tal classificação (projetos, pessoas e idéias) é identificar as diferentes formas de agrupamento e apontar suas principais características. Tais definições podem contribuir com pessoas, instituições ou empresas que pretendam utilizar o ciberespaço como ambiente de geração de conhecimento ou de desenvolvimento de atividades ou simplesmente troca de idéias. A partir das características de cada tipo de comunidade é possível se definir qual a melhor estrutura, levando-se em conta o interesse e objetivo de quem estrutura a comunidade ou a rede