A general theoretical framework for the design of artificial emotion systems in Autonomous Agents

Autonomous Agents (AAs) capable of exhibiting emotional behaviors have contributed to the development of natural human-machine interactions in several application domains. In order to provide AAs with emotional mechanisms, their underlying architecture must implement an Artificial Emotion System (AES), a computational model that imitates specific facets of human emotions. Although several AES have been reported in related literature, their design is generally supported on several emotion theories, leading researchers to model and integrate isolated emotion components and mechanisms into the architectures of AES. This theoretical foundation of AES contributes to ambiguities in the analysis and comparison of their underlying architectures, which demands the definition of standards, design guidelines, and integrative frameworks. In this paper, we present a psychologically inspired theoretical framework designed to serve as a platform for the unification of AES' components, the comparison of AES, and the design and implementation of AES in AAs. We analyze common emotion-related requirements of AES, emotion components involved in the design of this type of computational model, and emotion theories that drive the design of most AES. The validation of this framework demonstrates its compatibility with current AES and its feasibility as a model for unifying multiple emotional theories.     

A probabilistic computational model of cross-situational word learning

Words are the essence of communication: They are the building blocks of any language. Learning the meaning of words is thus one of the most important aspects of language acquisition: Children must first learn words before they can combine them into complex utterances. Many theories have been developed to explain the impressive efficiency of young children in acquiring the vocabulary of their language, as well as the developmental patterns observed in the course of lexical acquisition. A major source of disagreement among the different theories is whether children are equipped with special mechanisms and biases for word learning, or their general cognitive abilities are adequate for the task. We present a novel computational model of early word learning to shed light on the mechanisms that might be at work in this process. The model learns word meanings as probabilistic associations between words and semantic elements, using an incremental and probabilistic learning mechanism, and drawing only on general cognitive abilities. The results presented here demonstrate that much about word meanings can be learned from naturally occurring child-directed utterances (paired with meaning representations), without using any special biases or constraints, and without any explicit developmental changes in the underlying learning mechanism. Furthermore, our model provides explanations for the occasionally contradictory child experimental data, and offers predictions for the behavior of young word learners in novel situations.     

Mechanisms of developmental change in infant categorization

Computational models are tools for testing mechanistic theories of learning and development. Formal models allow us to instantiate theories of cognitive development in computer simulations. Model behavior can then be compared to real performance. Connectionist models, loosely based on neural information processing, have been successful in capturing a range of developmental phenomena, in particular on-line within-task category learning by young infants. Here we describe two new models. One demonstrates how age dependent changes in neural receptive field sizes can explain observed changes in on-line category learning between 3 and 10 months of age. The other aims to reconcile two conflicting views of infant categorization by focusing on the different task requirements of preferential looking and manual exploration studies. A dual-memory hypothesis posits that within-task category learning that drives looking time behaviors is based on a fast-learning memory system, whereas categorization based on background experience and assessed by paradigms requiring complex motor behavior relies on a second, slow-learning system. The models demonstrate how emphasizing the mechanistic causes of behaviors leads to discovery of deeper, more explanatory accounts of learning and development.     

Autonomy in Action: Linking the Act of Looking to Memory Formation in Infancy via Dynamic Neural Fields

Looking is a fundamental exploratory behavior by which infants acquire knowledge about the world. In theories of infant habituation, however, looking as an exploratory behavior has been deemphasized relative to the reliable nature with which looking indexes active cognitive processing. We present a new theory that connects looking to the dynamics of memory formation and formally implement this theory in a Dynamic Neural Field model that learns autonomously as it actively looks and looks away from a stimulus. We situate this model in a habituation task and illustrate the mechanisms by which looking, encoding, working memory formation, and long‐term memory formation give rise to habituation across multiple stimulus and task contexts. We also illustrate how the act of looking and the temporal dynamics of learning affect each other. Finally, we test a new hypothesis about the sources of developmental differences in looking.     

A Model of Knower-Level Behavior in Number Concept Development

Abstract We develop and evaluate a model of behavior on the Give- N task, a commonly used measure of young children's number knowledge. Our model uses the knower-level theory of how children represent numbers. To produce behavior on the Give- N task, the model assumes that children start out with a base rate that makes some answers more likely a priori than others but is updated on each experimental trial in a way that depends on the interaction between the experimenter's request and the child's knower level. We formalize this process as a generative graphical model, so that the parameters—including the base rate distribution and each child's knower level—can be inferred from data using Bayesian methods. Using this approach, we evaluate the model on previously published data from 82 children spanning the whole developmental range. The model provides an excellent fit to these data, and the inferences about the base rate and knower levels are interpretable and insightful. We discuss how our modeling approach can be extended to other developmental tasks and can be used to help evaluate alternative theories of number representation against the knower-level theory.     

Neonatal exposure to novel environment enhances hippocampal-dependent memory function during infancy and adulthood

Early life experience affects behavior and brain mechanisms. Handling rats during the first three weeks in life can slow age-related cognitive decline (as measured by a hippocampal-dependent spatial learning task) and reduce age-related hippocampal neuron loss. It is not clear, however, whether this early environmental influence on learning is selective for old age or is more general, affecting cognitive development during infancy and young adulthood as well. We briefly exposed neonatal rats to a novel non-home environment for 3 min daily during the first three weeks of life (as a component of the handling method). We found that this brief early environmental manipulation resulted in enhanced hippocampal-dependent learning immediately after weaning and that this learning enhancement persisted into adulthood. These results suggest that subtle early life events can affect cognitive development in all developmental stages and that changes in neural mechanisms other than neuron number are likely to mediate the learning enhancement at multiple developmental stages.     

Memory for object location: a span study in children.

The aim of the present study was to analyze the developmental changes in three spatial processes, namely, in positional reconstruction involving the retention of spatial locations per se (Positional encoding task), in the assignment of objects to positions (Object-to-position assignment task), and in the integration of these two (Combined task). A span procedure was used to assess the development of spatial memory in children aged 6, 8, and 10 years tested in these three tasks. The findings of the present study provide developmental spans for each relocation task. Results show an age-dependent improvement in all tasks, suggesting that spatial position is not automatically encoded. The results also show different developmental patterns for the relocation tasks considered, suggesting that spatial memory comprises a number of different component processes.     

Number-knower levels in young children: insights from Bayesian modeling

Lee and Sarnecka (2010) developed a Bayesian model of young children’s behavior on the Give-N test of number knowledge. This paper presents two new extensions of the model, and applies the model to new data. In the first extension, the model is used to evaluate competing theories about the conceptual knowledge underlying children’s behavior. One, the knower-levels theory, is basically a “stage” theory involving real conceptual change. The other, the approximate-meanings theory, assumes that the child’s conceptual knowledge is relatively constant, although performance improves over time. In the second extension, the model is used to ask whether the same latent psychological variable (a child’s number-knower level) can simultaneously account for behavior on two tasks (the Give-N task and the Fast-Cards task) with different performance demands. Together, these two demonstrations show the potential of the Bayesian modeling approach to improve our understanding of the development of human cognition.     

Generality with specificity: the dynamic field theory generalizes across tasks and time scales

A central goal in cognitive and developmental science is to develop models of behavior that can generalize across both tasks and development while maintaining a commitment to detailed behavioral prediction. This paper presents tests of one such model, the Dynamic Field Theory (DFT). The DFT was originally proposed to capture delay-dependent biases in spatial recall and developmental changes in spatial recall performance. More recently, the theory was generalized to adults' performance in a second spatial working memory task, position discrimination. Here we use the theory to predict a specific, complex developmental pattern in position discrimination. Data with 3- to 6-year-old children and adults confirm these predictions, demonstrating that the DFT achieves generality across tasks and time scales, as well as the specificity necessary to generate novel, falsifiable predictions.     

Personality predicts working-memory-related activation in the caudal anterior cingulate cortex

Behavioral studies suggest that two affective dimensions of personality are associated with working memory (WM) function. WM load is known to modulate neural activity in the caudal anterior cingulate cortex (ACC), a brain region critical for the cognitive control of behavior. On this basis, we hypothesized that neural activity in the caudal ACC during a WM task should be associated with personality: correlated negatively with behavioral approach sensitivity (BAS) and positively with behavioral inhibition sensitivity (BIS). Using functional magnetic resonance imaging, we measured brain activity in 14 participants performing a three-back WM task. Higher self-reported BAS predicted better WM performance (r = .27) and lower WM-related activation in the caudal ACC (r = -.84), suggesting personality differences in cognitive control. The data bolster approach-withdrawal (action control) theories of personality and suggest refinements to the dominant views of ACC and personality.     

A developmental perspective on behavioral determinants

This paper presents a developmental perspective on behavioral determinants which integrates some aspects of the Piagetian and Gibsonian theories as well as some basic postulates of AI theories. This developmental model is summarized in a series of assertions and illustrated by examples of various kinds of behavior. It assumes that three qualitatively different changes in behavioral determinants take place between birth and 18 years; this assumption means that the link between sensory information and action undergoes several transformations during development. Moreover, within each stage, behavior is determined by the degree of development of the subject's representation of objects and of himself. The initial state of behavior at birth is particularly discussed in this paper, with regard to imitation ability and auditory-visual coordination. The issue is whether or not development begins from a differentiated or non-differentiated subject-object relationship.     

The cost of simplifying complex developmental phenomena: a new perspective on learning to walk

Researchers can study complex developmental phenomena with all the inherent noise and complexity or simplify behaviors to hone in on the essential aspects of a phenomenon. We used the development of walking as a model system to compare the costs and benefits of simplifying a complex, noisy behavior. Traditionally, researchers simplify infant walking by recording gait measures as infants take continuous, forward steps along straight paths. Here, we compared the traditional straight‐path task with spontaneous walking during 20 minutes of free play in 97 infants (10.75–19.99 months of age). We recorded infants’ footfalls on an instrumented floor to calculate gait measures in the straight‐path and free‐play tasks. In addition, we scored videos for other critical aspects of spontaneous walking—steps per bout, shape of walking path, and step direction. Studying infant walking during free play incurred no cost compared with the straight‐path task, but considerable benefits. Straight‐path gait was highly correlated with spontaneous gait and both sets of measures improved with walking age, validating use of the straight‐path task as an index of development. However, a large proportion of free‐play bouts were too short to permit standard gait measures, and most bouts were curved with omnidirectional steps. The high prevalence of these “non‐canonical” bouts was constant over development. We propose that a focus on spontaneous walking, the phenomenon we ostensibly wish to explain, yields important insights into the problems infants solve while learning to walk. Other areas of developmental research may also benefit from retaining the complexity of complex phenomena.     

个体助人行为的形成与发展——基于扎根理论的探究

为了解个体助人行为的发生、发展和变化,本研究采用质性研究方法,围绕日常生活中助人行为深度访谈了10名始成年期的高助人者,运用扎根理论对访谈资料进行分析和理论建构。发现个体助人行为的发展可分成三个阶段：形成期、冲击期及稳固期,各发展阶段以对助人行为的认知和主要动机为核心,呈现出稳定-外控、不稳定-外控、稳定-内控的特点,据此提出助人行为的发展历程,并讨论了研究的理论价值、现实意义及局限。     

任务分析与教师的教学技能成长

本文结合笔者多年从事教师培训的经验,提出在我国课堂教学设计中需要引进并逐渐推广任务分析技术。其目的是引导教师学习现代学习论和教学论,使其教学行为受现代学习与教学心理学原理的支配,加快专业技能的成长。为此本文先介绍任务分析技术的起源与发展,然后论述任务分析对促进教师的教学技能成长的作用,最后用我们的教学研究结果对上述观点提供经验的支持。     

Ignorance or integration: the cognitive processes underlying choice behavior

The fast‐and‐frugal heuristic framework assumes noncompensatory tools for human preferences (i.e., priority heuristic) and inferences (i.e., take the best heuristic). According to this framework, these heuristics predict choice behavior as well as model the cognitive processes underlying such behavior. The current paper presents two studies that juxtapose predictions derived from these two simple heuristics with alternative predictions derived from compensatory principles. Dependent measures that included reaction time, choice pattern, confidence level, and accuracy were better predicted by compensatory indices than by noncompensatory indices. These findings suggest that people do not rely on limited arguments only, but tend to integrate all acquired information into their choice processes. This tendency was replicated even when the experimental task facilitated the use of noncompensatory principles. We argue that the fast and frugal heuristics can predict the final outcome only under certain conditions, but even in these particular situations they are not applicable to the processes underlying choice behavior. An integrative model for choice behavior is proposed that better represents the data. Copyright © 2009 John Wiley & Sons, Ltd.     

Young infants reach correctly in A-not-B tasks: on the development of stability and perseveration

This paper examines the development of perseverative reaching in the A-not-B task. We describe two recent models that view perseveration as a sign of developmental progress toward stability. In Experiment 1, we test the novel prediction from both models that very young infants should not perseverate in the A-not-B task whereas older infants should. We tracked infants' behavior monthly on the A-not-B task and found that infants reached correctly at 5 months, and only perseverated at 7 and 8 months of age. Experiment 2 provides further evidence on the role of motor development in the emergence of perseveration by exploring the connection between perseveration and detailed changes in reach kinematics in two infants across the first year. These data together suggest that perseveration is a sign of developmental achievement on the path to stable and flexible behavior.     

A connectionist modeling study of the neural mechanisms underlying pain’s ability to reorient attention

Connectionist modeling was used to investigate the brain mechanisms responsible for pain’s ability to shift attention away from another stimulus modality and toward itself. Different connectionist model architectures were used to simulate the different possible brain mechanisms underlying this attentional bias, where nodes in the model simulated the brain areas thought to mediate the attentional bias, and the connections between the nodes simulated the interactions between the brain areas. Mathematical optimization techniques were used to find the model parameters, such as connection strengths, that produced the best quantitative fits of reaction time and event-related potential data obtained in our previous work. Of the several architectures tested, two produced excellent quantitative fits of the experimental data. One involved an unexpected pain stimulus activating somatic threat detectors in the dorsal posterior insula. This threat detector activity was monitored by the medial prefrontal cortex, which in turn evoked a phasic response in the locus coeruleus. The locus coeruleus phasic response resulted in a facilitation of the cortical areas involved in decision and response processes time-locked to the painful stimulus. The second architecture involved the presence of pain causing an increase in general arousal. The increase in arousal was mediated by locus coeruleus tonic activity, which facilitated responses in the cortical areas mediating the sensory, decision, and response processes involved in the task. These two neural network architectures generated competing predictions that can be tested in future studies.     

Sampling Development

Research in developmental psychology requires sampling at different time points. Accurate depictions of developmental change provide a foundation for further empirical studies and theories about developmental mechanisms. However, overreliance on widely spaced sampling intervals in cross-sectional and longitudinal designs threatens the validity of the enterprise. This article discusses how to sample development in order to accurately discern the shape of developmental change. The ideal solution is daunting: to summarize behavior over 24-hour intervals and collect daily samples over the critical periods of change. We discuss the magnitude of errors due to undersampling, and the risks associated with oversampling. When daily sampling is not feasible, we offer suggestions for sampling methods that can provide preliminary reference points and provisional sketches of the general shape of a developmental trajectory. Denser sampling then can be applied strategically during periods of enhanced variability, inflections in the rate of developmental change, or in relation to key events or processes that may affect the course of change. Despite the challenges of dense repeated sampling, researchers must take seriously the problem of sampling on a developmental time scale if we are to know the true shape of developmental change.