Development of a Cultural Algorithm and a Trans-disciplinary Framework for Evolutionary Processes

随着时间的推移，文化元素会自适应地建立在彼此的基础上，变得多样化，并变得更加复杂。有人提出，文化就像早期的微生物生命一样，是通过公共交换的方式进化的，而不是优胜劣汰。这项研究计划旨在发展一个包含这些过程的对进化的一般理解。 这项研究将部分建立在EVOC的基础上，EVOC是一种基于公共交换的文化进化计算模型，由基于神经网络的代理组成，这些代理发明和模仿想法，从而进化出更适合和更多样化的行动。新的代理人将使用现有的概念相互作用数学模型来重组信息，该模型考虑了(1)概念在不同环境中如何变化，以及(2)人类如何根据情况在不同的思维模式之间转换。EVOC2将产生涉及商品发明和废物回收的经济网络。废物回收是EVOC2的理想应用，因为它需要将信息从一个领域重组到另一个领域。预计EVOC2将发展出不仅是累积性、适应性和开放式的新颖性，而且还具有变革性。EVOC2的框架将被抽象出来，以开发一种通用的文化算法(CA)，类似于遗传算法(GA)，这是一个具有不同应用的优化工具。在需要从不同角度考虑信息并因此进行重组的任务上，CA有望超过GA。 这项研究将以之前的努力为基础，为进化过程开发一个综合框架，包括文化进化和自然选择，以及早期和微生物生命。拟议的研究对于实现社会和行为科学的统一是必要的，可以与达尔文对生命科学的统一相媲美。预计它将促进对当前加快的文化变革步伐及其对人类和我们的星球的变革影响的科学理解，并帮助我们找到解决我们现在面临的复杂危机的办法。

Elements of culture adaptively build on one another, diversify, and become more complex over time. It has been proposed that culture, like early and microbial life, evolves through, not survival of the fittest, but transformation by way of communal exchange. This research program aims to develop a general understanding of evolution that incorporates these processes. The research will in part build on EVOC, a communal exchange based computational model of cultural evolution consisting of neural network based agents that invent and imitate ideas and thereby evolve fitter and more diverse actions. The new agents will restructure information using an existing mathematical model of concept interactions that accounts for (1) how concepts change in different contexts, and (2) how humans shift between different modes of thought depending on the situation. EVOC2 will generate economic networks involving both invention of goods and recycling of wastes. Waste recycling is an ideal application for EVOC2 because it entails restructuring of information from one domain to another. EVOC2 is expected to evolve novelty that is not just cumulative, adaptive, and open-ended, but transformative. The skeleton of EVOC2 will be abstracted to develop a general-purpose cultural algorithm (CA) analogous to the genetic algorithm (GA), an optimization tool with diverse applications. The CA is expected to outperform the GA on tasks that require that information be considered from a different perspective, and thereby restructured. The research will build on previous efforts toward the development of an integrated framework for evolutionary processes that encompasses cultural evolution and natural selection, as well as early and microbial life. The proposed research is necessary to achieve a unification of the social and behavioral sciences comparable with Darwin's unification of the life sciences. It is expected to promote a scientific understanding of the current accelerated pace of cultural change and its transformative effects on humans and our planet, and to aid us in finding solutions to complex crises we now face.