Activity in inferior parietal and medial prefrontal cortex signals the accumulation of evidence in a probability learning task.

In an uncertain environment, probabilities are key to predicting future events and making adaptive choices. However, little is known about how humans learn such probabilities and where and how they are encoded in the brain, especially when they concern more than two outcomes. During functional magnetic resonance imaging (fMRI), young adults learned the probabilities of uncertain stimuli through repetitive sampling. Stimuli represented payoffs and participants had to predict their occurrence to maximize their earnings. Choices indicated loss and risk aversion but unbiased estimation of probabilities. BOLD response in medial prefrontal cortex and angular gyri increased linearly with the probability of the currently observed stimulus, untainted by its value. Connectivity analyses during rest and task revealed that these regions belonged to the default mode network. The activation of past outcomes in memory is evoked as a possible mechanism to explain the engagement of the default mode network in probability learning. A BOLD response relating to value was detected only at decision time, mainly in striatum. It is concluded that activity in inferior parietal and medial prefrontal cortex reflects the amount of evidence accumulated in favor of competing and uncertain outcomes.

Machine learning algorithms for spatial data. Case studies: Environmental pollution, natural hazards, renewable resources

This paper presents general problems and approaches for the spatial data analysis using machine learning algorithms. Machine learning is a very powerful approach to adaptive data analysis, modelling and visualisation. The key feature of the machine learning algorithms is that they learn from empirical data and can be used in cases when the modelled environmental phenomena are hidden, nonlinear, noisy and highly variable in space and in time. Most of the machines learning algorithms are universal and adaptive modelling tools developed to solve basic problems of learning from data: classification/pattern recognition, regression/mapping and probability density modelling. In the present report some of the widely used machine learning algorithms, namely artificial neural networks (ANN) of different architectures and Support Vector Machines (SVM), are adapted to the problems of the analysis and modelling of geo-spatial data. Machine learning algorithms have an important advantage over traditional models of spatial statistics when problems are considered in a high dimensional geo-feature spaces, when the dimension of space exceeds 5. Such features are usually generated, for example, from digital elevation models, remote sensing images, etc. An important extension of models concerns considering of real space constrains like geomorphology, networks, and other natural structures. Recent developments in semi-supervised learning can improve modelling of environmental phenomena taking into account on geo-manifolds. An important part of the study deals with the analysis of relevant variables and models' inputs. This problem is approached by using different feature selection/feature extraction nonlinear tools. To demonstrate the application of machine learning algorithms several interesting case studies are considered: digital soil mapping using SVM, automatic mapping of soil and water system pollution using ANN; natural hazards risk analysis (avalanches, landslides), assessments of renewable resources (wind fields) with SVM and ANN models, etc. The dimensionality of spaces considered varies from 2 to more than 30. Figures 1, 2, 3 demonstrate some results of the studies and their outputs. Finally, the results of environmental mapping are discussed and compared with traditional models of geostatistics.

Coevolution of adaptive technology, maladaptive culture and population size in a producer-scrounger game.

Technology (i.e. tools, methods of cultivation and domestication, systems of construction and appropriation, machines) has increased the vital rates of humans, and is one of the defining features of the transition from Malthusian ecological stagnation to a potentially perpetual rising population growth. Maladaptations, on the other hand, encompass behaviours, customs and practices that decrease the vital rates of individuals. Technology and maladaptations are part of the total stock of culture carried by the individuals in a population. Here, we develop a quantitative model for the coevolution of cumulative adaptive technology and maladaptive culture in a 'producer-scrounger' game, which can also usefully be interpreted as an 'individual-social' learner interaction. Producers (individual learners) are assumed to invent new adaptations and maladaptations by trial-and-error learning, insight or deduction, and they pay the cost of innovation. Scroungers (social learners) are assumed to copy or imitate (cultural transmission) both the adaptations and maladaptations generated by producers. We show that the coevolutionary dynamics of producers and scroungers in the presence of cultural transmission can have a variety of effects on population carrying capacity. From stable polymorphism, where scroungers bring an advantage to the population (increase in carrying capacity), to periodic cycling, where scroungers decrease carrying capacity, we find that selection-driven cultural innovation and transmission may send a population on the path of indefinite growth or to extinction.

Efficient learning of reactive robot behaviors with a Neural-Q_learning approach

The purpose of this paper is to propose a Neural-Q_learning approach designed for online learning of simple and reactive robot behaviors. In this approach, the Q_function is generalized by a multi-layer neural network allowing the use of continuous states and actions. The algorithm uses a database of the most recent learning samples to accelerate and guarantee the convergence. Each Neural-Q_learning function represents an independent, reactive and adaptive behavior which maps sensorial states to robot control actions. A group of these behaviors constitutes a reactive control scheme designed to fulfill simple missions. The paper centers on the description of the Neural-Q_learning based behaviors showing their performance with an underwater robot in a target following task. Real experiments demonstrate the convergence and stability of the learning system, pointing out its suitability for online robot learning. Advantages and limitations are discussed

On optimal learning schedules and the marginal value of cumulative cultural evolution.

The age-dependent choice between expressing individual learning (IL) or social learning (SL) affects cumulative cultural evolution. A learning schedule in which SL precedes IL is supportive of cumulative culture because the amount of nongenetically encoded adaptive information acquired by previous generations can be absorbed by an individual and augmented. Devoting time and energy to learning, however, reduces the resources available for other life-history components. Learning schedules and life history thus coevolve. Here, we analyze a model where individuals may have up to three distinct life stages: "infants" using IL or oblique SL, "juveniles" implementing IL or horizontal SL, and adults obtaining material resources with learned information. We study the dynamic allocation of IL and SL within life stages and how this coevolves with the length of the learning stages. Although no learning may be evolutionary stable, we find conditions where cumulative cultural evolution can be selected for. In that case, the evolutionary stable learning schedule causes individuals to use oblique SL during infancy and a mixture between IL and horizontal SL when juvenile. We also find that the selected pattern of oblique SL increases the amount of information in the population, but horizontal SL does not do so.

Adaptive search and information updating in sequential mate choice

Classical treatments of problems of sequential mate choice assume that the distribution of the quality of potential mates is known a priori. This assumption, made for analytical purposes, may seem unrealistic, opposing empirical data as well as evolutionary arguments. Using stochastic dynamic programming, we develop a model that includes the possibility for searching individuals to learn about the distribution and in particular to update mean and variance during the search. In a constant environment, a priori knowledge of the parameter values brings strong benefits in both time needed to make a decision and average value of mate obtained. Knowing the variance yields more benefits than knowing the mean, and benefits increase with variance. However, the costs of learning become progressively lower as more time is available for choice. When parameter values differ between demes and/or searching periods, a strategy relying on fixed a priori information might lead to erroneous decisions, which confers advantages on the learning strategy. However, time for choice plays an important role as well: if a decision must be made rapidly, a fixed strategy may do better even when the fixed image does not coincide with the local parameter values. These results help in delineating the ecological-behavior context in which learning strategies may spread.

Adaptive integral assessment package for the A2UN@ project

This paper presents a first approach of Evaluation Engine Architecture (EEA) as proposal to support adaptive integral assessment, in the context of a virtual learning environment. The goal of our research is design an evaluation engine tool to assist in the whole assessment process within the A2UN@ project, linking that tool with the other key elements of a learning design (learning task, learning resources and learning support). The teachers would define the relation between knowledge, competencies, activities, resources and type of assessment. Providing this relation is possible obtain more accurate estimations of student's knowledge for adaptive evaluations and future recommendations. The process is supported by usage of educational standards and specifications and for an integral user modelling

Searching for and positioning of contextualized learning objects

Learning object economies are marketplaces for the sharing and reuse of learning objects (LO). There are many motivations for stimulating the development of the LO economy. The main reason is the possibility of providing the right content, at the right time, to the right learner according to adequate quality standards in the context of a lifelong learning process; in fact, this is also the main objective of education. However, some barriers to the development of a LO economy, such as the granularity and editability of LO, must be overcome. Furthermore, some enablers, such as learning design generation and standards usage, must be promoted in order to enhance LO economy. For this article, we introduced the integration of distributed learning object repositories (DLOR) as sources of LO that could be placed in adaptive learning designs to assist teachers’ design work. Two main issues presented as a result: how to access distributed LO, and where to place the LO in the learning design. To address these issues, we introduced two processes: LORSE, a distributed LO searching process, and LOOK, a micro context-based positioning process, respectively. Using these processes, the teachers were able to reuse LO from different sources to semi-automatically generate an adaptive learning design without leaving their virtual environment. A layered evaluation yielded good results for the process of placing learning objects from controlled learning object repositories into a learning design, and permitting educators to define different open issues that must be covered when they use uncontrolled learning object repositories for this purpose. We verified the satisfaction users had with our solution

Learning ability and longevity: a symmetrical evolutionary trade-off in Drosophila.

Learning ability can be substantially improved by artificial selection in animals ranging from Drosophila to rats. Thus these species have not used their evolutionary potential with respect to learning ability, despite intuitively expected and experimentally demonstrated adaptive advantages of learning. This suggests that learning is costly, but this notion has rarely been tested. Here we report correlated responses of life-history traits to selection for improved learning in Drosophila melanogaster. Replicate populations selected for improved learning lived on average 15% shorter than the corresponding unselected control populations. They also showed a minor reduction in fecundity late in life and possibly a minor increase in dry adult mass. Selection for improved learning had no effect on egg-to-adult viability, development rate, or desiccation resistance. Because shortened longevity was the strongest correlated response to selection for improved learning, we also measured learning ability in another set of replicate populations that had been selected for extended longevity. In a classical olfactory conditioning assay, these long-lived flies showed an almost 40% reduction in learning ability early in life. This effect disappeared with age. Our results suggest a symmetrical evolutionary trade-off between learning ability and longevity in Drosophila.

Rational reasoning or adaptive behavior? Evidence from two-person beauty contest games

Many experiments have shown that human subjects do not necessarily behave in line with game theoretic assumptions and solution concepts. The reasons for this non-conformity are multiple. In this paper we study the argument whether a deviation from game theory is because subjects are rational, but doubt that others are rational as well, compared to the argument that subjects, in general, are boundedly rational themselves. To distinguish these two hypotheses, we study behavior in repeated 2-person and many-person Beauty-Contest-Games which are strategically different from one another. We analyze four different treatments and observe that convergence toward equilibrium is driven by learning through the information about the other player s choice and adaptation rather than self-initiated rational reasoning.

Influence of learning on range expansion and adaptation to novel habitats.

Learning has been postulated to 'drive' evolution, but its influence on adaptive evolution in heterogeneous environments has not been formally examined. We used a spatially explicit individual-based model to study the effect of learning on the expansion and adaptation of a species to a novel habitat. Fitness was mediated by a behavioural trait (resource preference), which in turn was determined by both the genotype and learning. Our findings indicate that learning substantially increases the range of parameters under which the species expands and adapts to the novel habitat, particularly if the two habitats are separated by a sharp ecotone (rather than a gradient). However, for a broad range of parameters, learning reduces the degree of genetically-based local adaptation following the expansion and facilitates maintenance of genetic variation within local populations. Thus, in heterogeneous environments learning may facilitate evolutionary range expansions and maintenance of the potential of local populations to respond to subsequent environmental changes.

Statistical learning theory for geospatial data. Case study: Aral sea

In recent years there has been an explosive growth in the development of adaptive and data driven methods. One of the efficient and data-driven approaches is based on statistical learning theory (Vapnik 1998). The theory is based on Structural Risk Minimisation (SRM) principle and has a solid statistical background. When applying SRM we are trying not only to reduce training error ? to fit the available data with a model, but also to reduce the complexity of the model and to reduce generalisation error. Many nonlinear learning procedures recently developed in neural networks and statistics can be understood and interpreted in terms of the structural risk minimisation inductive principle. A recent methodology based on SRM is called Support Vector Machines (SVM). At present SLT is still under intensive development and SVM find new areas of application (www.kernel-machines.org). SVM develop robust and non linear data models with excellent generalisation abilities that is very important both for monitoring and forecasting. SVM are extremely good when input space is high dimensional and training data set i not big enough to develop corresponding nonlinear model. Moreover, SVM use only support vectors to derive decision boundaries. It opens a way to sampling optimization, estimation of noise in data, quantification of data redundancy etc. Presentation of SVM for spatially distributed data is given in (Kanevski and Maignan 2004).

Insights from insects about adaptive social information use.

Copying others can greatly improve individual fitness and is fundamental for the organisation of societies. Yet in some situations it is better to ignore social information and either explore the world individually or use personal information obtained through prior experience. Insects provide excellent models to study the strategic use of social information, but insights from recent research have rarely been viewed in the light of social learning strategies. Here we discuss how insects tailor their reliance on social information to those circumstances for which it is most beneficial, and suggest that insects and vertebrates use similar information-use strategies. We highlight future research avenues, including the use of molecular tools to study the genetic and genomic basis of social information use.

The underestimation of compliance to direct help requests explained by adaptive sampling

Helping behavior is any intentional behavior that benefits another living being or group (Hogg & Vaughan, 2010). People tend to underestimate the probability that others will comply with their direct requests for help (Flynn & Lake, 2008). This implies that when they need help, they will assess the probability of getting it (De Paulo, 1982, cited in Flynn & Lake, 2008) and then they will tend to estimate one that is actually lower than the real chance, so they may not even consider worth asking for it. Existing explanations for this phenomenon attribute it to a mistaken cost computation by the help seeker, who will emphasize the instrumental cost of “saying yes”, ignoring that the potential helper also needs to take into account the social cost of saying “no”. And the truth is that, especially in face-to-face interactions, the discomfort caused by refusing to help can be very high. In short, help seekers tend to fail to realize that it might be more costly to refuse to comply with a help request rather than accepting. A similar effect has been observed when estimating trustworthiness of people. Fetchenhauer and Dunning (2010) showed that people also tend to underestimate it. This bias is reduced when, instead of asymmetric feedback (getting feedback only when deciding to trust the other person), symmetric feedback (always given) was provided. This cause could as well be applicable to help seeking as people only receive feedback when they actually make their request but not otherwise. Fazio, Shook, and Eiser (2004) studied something that could be reinforcing these outcomes: Learning asymmetries. By means of a computer game called BeanFest, they showed that people learn better about negatively valenced objects (beans in this case) than about positively valenced ones. This learning asymmetry esteemed from “information gain being contingent on approach behavior” (p. 293), which could be identified with what Fetchenhauer and Dunning mention as ‘asymmetric feedback’, and hence also with help requests. Fazio et al. also found a generalization asymmetry in favor of negative attitudes versus positive ones. They attributed it to a negativity bias that “weights resemblance to a known negative more heavily than resemblance to a positive” (p. 300). Applied to help seeking scenarios, this would mean that when facing an unknown situation, people would tend to generalize and infer that is more likely that they get a negative rather than a positive outcome from it, so, along with what it was said before, people will be more inclined to think that they will get a “no” when requesting help. Denrell and Le Mens (2011) present a different perspective when trying to explain judgment biases in general. They deviate from the classical inappropriate information processing (depicted among other by Fiske & Taylor, 2007, and Tversky & Kahneman, 1974) and explain this in terms of ‘adaptive sampling’. Adaptive sampling is a sampling mechanism in which the selection of sample items is conditioned by the values of the variable of interest previously observed (Thompson, 2011). Sampling adaptively allows individuals to safeguard themselves from experiences they went through once and turned out to lay negative outcomes. However, it also prevents them from giving a second chance to those experiences to get an updated outcome that could maybe turn into a positive one, a more positive one, or just one that regresses to the mean, whatever direction that implies. That, as Denrell and Le Mens (2011) explained, makes sense: If you go to a restaurant, and you did not like the food, you do not choose that restaurant again. This is what we think could be happening when asking for help: When we get a “no”, we stop asking. And here, we want to provide a complementary explanation for the underestimation of the probability that others comply with our direct help requests based on adaptive sampling. First, we will develop and explain a model that represents the theory. Later on, we will test it empirically by means of experiments, and will elaborate on the analysis of its results.

Multidimensional Adaptations for Open Learning Management Systems

Our work is focused on alleviating the workload for designers of adaptive courses on the complexity task of authoring adaptive learning designs adjusted to specific user characteristics and the user context. We propose an adaptation platform that consists in a set of intelligent agents where each agent carries out an independent adaptation task. The agents apply machine learning techniques to support the user modelling for the adaptation process