A hierarchical method for finding optimal architecture and weights using evolutionary least square based learning

In this paper, we present a novel approach of implementing a combination methodology to find appropriate neural network architecture and weights using an evolutionary least square based algorithm (GALS).1 This paper focuses on aspects such as the heuristics of updating weights using an evolutionary least square based algorithm, finding the number of hidden neurons for a two layer feed forward neural network, the stopping criterion for the algorithm and finally some comparisons of the results with other existing methods for searching optimal or near optimal solution in the multidimensional complex search space comprising the architecture and the weight variables. We explain how the weight updating algorithm using evolutionary least square based approach can be combined with the growing architecture model to find the optimum number of hidden neurons. We also discuss the issues of finding a probabilistic solution space as a starting point for the least square method and address the problems involving fitness breaking. We apply the proposed approach to XOR problem, 10 bit odd parity problem and many real-world benchmark data sets such as handwriting data set from CEDAR, breast cancer and heart disease data sets from UCI ML repository. The comparative results based on classification accuracy and the time complexity are discussed.     

Toward a theory of evolutionary computation

We outline a theory of evolutionary computation using a formal model of evolutionary computation--the Evolutionary Turing Machine--which is introduced as the extension of the Turing Machine model. Evolutionary Turing Machines provide a better and a more complete model for evolutionary computing than conventional Turing Machines, algorithms, and Markov chains. The convergence and convergence rate are defined and investigated in terms of this new model. The sufficient conditions needed for the completeness and optimality of evolutionary search are investigated. In particular, the notion of the total optimality as an instance of the multiobjective optimization of the Universal Evolutionary Turing Machine is introduced. This provides an automatic way to deal with the intractability of evolutionary search by optimizing the quality of solutions and search costs simultaneously. Based on a new model a very flexible classification of optimization problem hardness for the evolutionary techniques is proposed. The expressiveness of evolutionary computation is investigated. We show that the problem of the best evolutionary algorithm is undecidable independently of whether the fitness function is time dependent or fixed. It is demonstrated that the evolutionary computation paradigm is more expressive than Turing Machines, and thus the conventional computer science based on them. We show that an Evolutionary Turing Machine is able to solve nonalgorithmically the halting problem of the Universal Turing Machine and, asymptotically, the best evolutionary algorithm problem. In other words, the best evolutionary algorithm does not exist, but it can be potentially indefinitely approximated using evolutionary techniques.     

The evolutionary psychology of sexual harassment in organizations

We explore the value of theory developed in the field of evolutionary psychology for predicting and explaining patterns of sociosexual interaction between men and women in the workplace. In particular, we focus on assessing the value of this theoretical framework for increasing our understanding of the behavioral phenomenon of sexual harassment in organizations. To do this, we develop hypotheses based on a consideration of the evolved psychological mechanisms which have been proposed by evolutionary psychologists to underly human sexual behavior. We then review the existing literature on sociosexual behavior and sexual harassment within the workplace as a preliminary test of predictions concerning the expected profile, reaction, and motivation of recipients and initiators of sexual advances in the workplace. We also analyze the expected effect of novel aspects of the organizational environment on overt sexual behavior. This theory predicts, for example, that most sexual harassment will be initiated by men, and that young, single women will be the most frequent victims. This theory also predicts that women will generally experience a more negative reaction than men to sexual advances in the workplace, but that this emotional response will be mediated by other factors such as the profile of the initiator. The available data generally confirm these and other related predictions. Moreover, these results are quite consistent with other studies of the evolutionary psychology of human sexuality. We discuss the implications for theory in this area, and suggest some potentially fruitful avenues for original empirical research on human sexual behavior, inside or outside the workplace.     

Developing the genotype-to-phenotype relationship in evolutionary theory: A primer of developmental features

For decades, there have been repeated calls for more integration across evolutionary and developmental biology. However, critiques in the literature and recent funding initiatives suggest this integration remains incomplete. We suggest one way forward is to consider how we elaborate the most basic concept of development, the relationship between genotype and phenotype, in traditional models of evolutionary processes. For some questions, when more complex features of development are accounted for, predictions of evolutionary processes shift. We present a primer on concepts of development to clarify confusion in the literature and fuel new questions and approaches. The basic features of development involve expanding a base model of genotype-to-phenotype to include the genome, space, and time. A layer of complexity is added by incorporating developmental systems, including signal-response systems and networks of interactions. The developmental emergence of function, which captures developmental feedbacks and phenotypic performance, offers further model elaborations that explicitly link fitness with developmental systems. Finally, developmental features such as plasticity and developmental niche construction conceptualize the link between a developing phenotype and the external environment, allowing for a fuller inclusion of ecology in evolutionary models. Incorporating aspects of developmental complexity into evolutionary models also accommodates a more pluralistic focus on the causal importance of developmental systems, individual organisms, or agents in generating evolutionary patterns. Thus, by laying out existing concepts of development, and considering how they are used across different fields, we can gain clarity in existing debates around the extended evolutionary synthesis and pursue new directions in evolutionary developmental biology. Finally, we consider how nesting developmental features in traditional models of evolution can highlight areas of evolutionary biology that need more theoretical attention.     

Environment changes epistasis to alter trade‐offs along alternative evolutionary paths

The fitness effect of a mutation can depend on both its genetic background, known as epistasis, and the prevailing external environment. Many examples of these dependencies are known, but few studies consider both aspects in combination, especially as they affect mutations that have been selected together. We examine interactions between five coevolved mutations in eight diverse environments. We find that mutations are, on average, beneficial across environments, but that there is high variation in their fitness effects, including many examples of mutations conferring a cost in some, but not other, genetic background‐environment combinations. Indeed, even when global interaction trends are accounted for, specific local mutation interactions are common and differed across environments. One consequence of this dependence is that the range of trade‐offs in genotype fitness across selected and alternative environments are contingent on the particular evolutionary path followed over the mutation landscape. Finally, although specific interactions were common, there was a consistent pattern of diminishing returns epistasis whereby mutation effects were less beneficial when added to genotypes of higher fitness. Our results underline that specific mutation effects are highly dependent on the combination of genetic and external environments, and support a general relationship between a genotype's current fitness and its potential to increase in fitness.     

Species interactions alter evolutionary responses to a novel environment

Studies of evolutionary responses to novel environments typically consider single species or perhaps pairs of interacting species. However, all organisms co-occur with many other species, resulting in evolutionary dynamics that might not match those predicted using single species approaches. Recent theories predict that species interactions in diverse systems can influence how component species evolve in response to environmental change. In turn, evolution might have consequences for ecosystem functioning. We used experimental communities of five bacterial species to show that species interactions have a major impact on adaptation to a novel environment in the laboratory. Species in communities diverged in their use of resources compared with the same species in monocultures and evolved to use waste products generated by other species. This generally led to a trade-off between adaptation to the abiotic and biotic components of the environment, such that species evolving in communities had lower growth rates when assayed in the absence of other species. Based on growth assays and on nuclear magnetic resonance (NMR) spectroscopy of resource use, all species evolved more in communities than they did in monocultures. The evolutionary changes had significant repercussions for the functioning of these experimental ecosystems: communities reassembled from isolates that had evolved in polyculture were more productive than those reassembled from isolates that had evolved in monoculture. Our results show that the way in which species adapt to new environments depends critically on the biotic environment of co-occurring species. Moreover, predicting how functioning of complex ecosystems will respond to an environmental change requires knowing how species interactions will evolve.     

Two different evolutionary origins of stem cell systems and their molecular basis

We propose two major evolutionary origins of stem cell systems in the animal kingdom. Adult pluripotent stem cell systems are found in many invertebrates and probably evolved as components of asexual reproduction. Lineage-specific stem cell systems probably evolved later and include neural and hematopoietic stem cell types. We propose that these two types of stem cell systems evolved independently. The vasa-like genes regulate reproductive stem cells, but not lineage-specific stem cells, which may be regulated by gcm genes. Here, we review the evidence for the molecular basis for the evolutionary origin of these two different stem cell systems.     

The arts and human nature: evolutionary aesthetics and the evolutionary status of art behaviours

This essay reviews one of the most recent books in a trend of new publications proffering evolutionary theorising about aesthetics and the arts—themes within an increasing literature on aspects of human life and human nature in terms of evolutionary theory. Stephen Davies’ The Artful Species links some of our aesthetic sensibilities with our evolved human nature and critically surveys the interdisciplinary debate regarding the evolutionary status of the arts. Davies’ engaging and accessible writing succeeds in demonstrating the maturity and scope of the field and his critique is timely and unparalleled. A laudable effort, however it may have benefited from espousing a co-evolutionary model more explicitly. Moreover there may be reason to question the usefulness of the standard set of distinctions (‘adaptation’, ‘spandrel’, ‘technology’) that Davies appeals to.     

Multitrophic effects of herbivore-induced plant volatiles in an evolutionary context

Herbivorous and carnivorous arthropods use plant volatiles when foraging for food. In response to herbivory, plants emit a blend that may be quantitatively and qualitatively different from the blend emitted when intact. This induced volatile blend alters the interactions of the plant with its environment. We review recent developments regarding the induction mechanism as well as the ecological consequences in a multitrophic and evolutionary context. It has been well established that carnivores (predators and parasitoids) are attracted by the volatiles induced by their herbivorous victims. This concerns an active plant response. In the case of attraction of predators, this is likely to result in a fitness benefit to the plant, because through consumption a predator removes the herbivores from the plant. However, the benefit to the plant is less clear when parasitoids are attracted, because parasitisation does usually not result in an instantaneous or in a complete termination of consumption by the herbivore. Recently, empirical evidence has been obtained that shows that the plant's response can increase plant fitness, in terms of seed production, due to a reduced consumption rate of parasitized herbivores. However, apart from a benefit from attracting carnivores, the induced volatiles can have a serious cost because there is an increasing number of studies that show that herbivores can be attracted. However, this does not necessarily result in settlement of the herbivores on the emitting plant. The presence of cues from herbivores and/or carnivores that indicate that the plant is a competitor- and/or enemy-dense space, may lead to an avoidance response. Thus, the benefit of emission of induced volatiles is likely to depend on the prevailing faunal composition. Whether plants can adjust their response and influence the emission of the induced volatiles, taking the prevalent environmental conditions into account, is an interesting question that needs to be addressed. The induced volatiles may also affect interactions of the emitting plant with its neighbours, e.g., through altered competitive ability or by the neighbour exploiting the emitted information.Major questions to be addressed in this research field comprise mechanistic aspects, such as the identification of the minimally effective blend of volatiles that explains the attraction of carnivores to herbivore-infested plants, and evolutionary aspects such as the fitness consequences of induced volatiles. The elucidation of mechanistic aspects is important for addressing ecological and evolutionary questions. For instance, an important tool to address ecological and evolutionary aspects would be to have plant pairs that differ in only a single trait. Such plants are likely to become available in the near future as a result of mechanistic studies on signal-transduction pathways and an increased interest in molecular genetics.     

Can possible evolutionary outcomes be determined directly from the population dynamics?

Traditionally, to determine the possible evolutionary behaviour of an ecological system using adaptive dynamics, it is necessary to calculate the fitness and its derivatives at a singular point. We investigate the claim that the possible evolutionary behaviour can be predicted directly from the population dynamics, without the need for calculation, by applying three criteria — one based on the form of the density dependent rates and two on the role played by the evolving parameters. Taking a general continuous time model, with broad ecological range, we show that the claim is true. Initially, we assume that individuals enter in class 1 and move through population classes sequentially; later we relax these assumptions and find that the criteria still apply. However, when we consider models where the evolving parameters appear non-linearly in the dynamics, we find some aspects of the criteria fail; useful but weaker results on possible evolutionary behaviour now apply.     

Sociality and health: impacts of sociality on disease susceptibility and transmission in animal and human societies

This paper introduces a theme issue presenting the latest developments in research on the impacts of sociality on health and fitness. The articles that follow cover research on societies ranging from insects to humans. Variation in measures of fitness (i.e. survival and reproduction) has been linked to various aspects of sociality in humans and animals alike, and variability in individual health and condition has been recognized as a key mediator of these relationships. Viewed from a broad evolutionary perspective, the evolutionary transitions from a solitary lifestyle to group living have resulted in several new health-related costs and benefits of sociality. Social transmission of parasites within groups represents a major cost of group living, but some behavioural mechanisms, such as grooming, have evolved repeatedly to reduce this cost. Group living also has created novel costs in terms of altered susceptibility to infectious and non-infectious disease as a result of the unavoidable physiological consequences of social competition and integration, which are partly alleviated by social buffering in some vertebrates. Here, we define the relevant aspects of sociality, summarize their health-related costs and benefits, and discuss possible fitness measures in different study systems. Given the pervasive effects of social factors on health and fitness, we propose a synthesis of existing conceptual approaches in disease ecology, ecological immunology and behavioural neurosciences by adding sociality as a key factor, with the goal to generate a broader framework for organismal integration of health-related research.     

Ecological aspects of the evolutionary processes

Darwin in his On the Origin of species made it clear that evolutionary change depends on the combined action of two different causes, the first being the origin of genetically based phenotypic variation in the individual organisms comprising the population and the second being the action of selective agents of the external environment placing demands on the individual organisms. For over a century following Darwin, most evolutionists focused on the origin of inherited variation and its transmission; many workers continue to regard genetics to be the core of evolutionary theory. Far less attention has been given to the exact nature of the selective agents with most evolutionists still treating this cause imprecisely to the detriment of our understanding of both nomological and historical evolutionary theory. Darwin was vague in the meaning of his new concept of "Natural Selection," using it interchangeably as one of the causes for evolutionary change and as the final outcome (= evolutionary change). In 1930, natural selection was defined clearly as "non-random, differential reproduction of genes" by R. Fisher and J.B.S. Haldane which is a statement of the outcome of evolutionary process and which omits mention of the causes bringing about this change. Evolutionists quickly accepted this outcome definition of natural selection, and have used interchangeably selection both as a cause and as the result of evolutionary change, causing great confusion. Herein, the details will be discussed of how the external environment (i.e., the environment-phenotype interaction) serves as selective agents and exerts demands on the phenotypic organisms. Included are the concepts of fitness and of the components of fitness (= adaptations) which are respectively (a) survival, (b) direct reproductive and (c) indirect reproductive features. Finally, it will be argued that historical-narrative analyses of organisms, including classification and phylogenetic history, are possible only with a full understanding of nomological evolutionary theory and with functional/adaptive studies of the employed taxonomic features in addition to the standard comparative investigations.     

Modelling evolutionary cell behaviour using neural networks: Application to tumour growth

In this paper, we present a modelling framework for cellular evolution that is based on the notion that a cell’s behaviour is driven by interactions with other cells and its immediate environment. We equip each cell with a phenotype that determines its behaviour and implement a decision mechanism to allow evolution of this phenotype. This decision mechanism is modelled using feed-forward neural networks, which have been suggested as suitable models of cell signalling pathways. The environmental variables are presented as inputs to the network and result in a response that corresponds to the phenotype of the cell. The response of the network is determined by the network parameters, which are subject to mutations when the cells divide. This approach is versatile as there are no restrictions on what the input or output nodes represent, they can be chosen to represent any environmental variables and behaviours that are of importance to the cell population under consideration. This framework was implemented in an individual-based model of solid tumour growth in order to investigate the impact of the tissue oxygen concentration on the growth and evolutionary dynamics of the tumour. Our results show that the oxygen concentration affects the tumour at the morphological level, but more importantly has a direct impact on the evolutionary dynamics. When the supply of oxygen is limited we observe a faster divergence away from the initial genotype, a higher population diversity and faster evolution towards aggressive phenotypes. The implementation of this framework suggests that this approach is well suited for modelling systems where evolution plays an important role and where a changing environment exerts selection pressure on the evolving population.     

Modeling social and evolutionary games

When game theory was introduced to biology, the components of classic game theory models were replaced with elements more befitting evolutionary phenomena. The actions of intelligent agents are replaced by phenotypic traits; utility is replaced by fitness; rational deliberation is replaced by natural selection. In this paper, I argue that this classic conception of comprehensive reapplication is misleading, for it overemphasizes the discontinuity between human behavior and evolved traits. Explicitly considering the representational roles of evolutionary game theory brings to attention areas of overlap that are often neglected, and so a range of evolutionary possibilities that are often overlooked. The clarifications this analysis provides are well illustrated by-and particularly valuable for-game theoretic treatments of the evolution of social behavior.     

Fitness consequences of variable maternal provisioning in quacking frogs (Crinia georgiana)

Variable maternal provisioning may evolve when there is variation in the quality of offspring environments. The frog Crinia georgiana has high variability in egg size both within and between clutches, independent of female phenotype. It breeds in ponds with high spatial and temporal variation in habitat quality. Egg size strongly affected offspring fitness in good and poor quality offspring environments, whether the egg size difference was from between or within clutches. Since there is a trade-off in egg size and number, these fitness consequences translate to strong effects on maternal fitness. In the variable and unpredictable offspring environment of C. georgiana, the maintenance of variable maternal provisioning both within and between clutches is likely to be an evolved response to the offspring environment.     

Phenotypic plasticity in evolutionary rescue experiments

Population persistence in a new and stressful environment can be influenced by the plastic phenotypic responses of individuals to this environment, and by the genetic evolution of plasticity itself. This process has recently been investigated theoretically, but testing the quantitative predictions in the wild is challenging because (i) there are usually not enough population replicates to deal with the stochasticity of the evolutionary process, (ii) environmental conditions are not controlled, and (iii) measuring selection and the inheritance of traits affecting fitness is difficult in natural populations. As an alternative, predictions from theory can be tested in the laboratory with controlled experiments. To illustrate the feasibility of this approach, we briefly review the literature on the experimental evolution of plasticity, and on evolutionary rescue in the laboratory, paying particular attention to differences and similarities between microbes and multicellular eukaryotes. We then highlight a set of questions that could be addressed using this framework, which would enable testing the robustness of theoretical predictions, and provide new insights into areas that have received little theoretical attention to date.     

Evolution of Dispersal in a Structured Metapopulation Model in Discrete Time

In this article, a structured metapopulation model in discrete time with catastrophes and density-dependent local growth is introduced. The fitness of a rare mutant in an environment set by the resident is defined, and an efficient method to calculate fitness is presented. With this fitness measure evolutionary analysis of this model becomes feasible. This article concentrates on the evolution of dispersal. The effect of catastrophes, dispersal cost, and local dynamics on the evolution of dispersal is investigated. It is proved that without catastrophes, if all population–dynamical attractors are fixed points, there will be selection for no dispersal. A new mechanism for evolutionary branching is also found: Even though local population sizes approach fixed points, catastrophes can cause enough temporal variability, so that evolutionary branching becomes possible.     

Reverse evolution of fitness in Drosophila melanogaster

Abstract The evolution of fitness is central to evolutionary theory, yet few experimental systems allow us to track its evolution in genetically and environmentally relevant contexts. Reverse evolution experiments allow the study of the evolutionary return to ancestral phenotypic states, including fitness. This in turn permits well‐defined tests for the dependence of adaptation on evolutionary history and environmental conditions. In the experiments described here, 20 populations of heterogeneous evolutionary histories were returned to their common ancestral environment for 50 generations, and were then compared with both their immediate differentiated ancestors and populations which had remained in the ancestral environment. One measure of fitness returned to ancestral levels to a greater extent than other characters did. The phenotypic effects of reverse evolution were also contingent on previous selective history. Moreover, convergence to the ancestral state was highly sensitive to environmental conditions. The phenotypic plasticity of fecundity, a character directly selected for, evolved during the experimental time frame. Reverse evolution appears to force multiple, diverged populations to converge on a common fitness state through different life‐history and genetic changes.