This paper summarizes recent research on competition-based learning procedures performed by the Navy Center for Applied Research in Artificial Intelligence at the Naval Research Laboratory. We have focused on a particularly interesting class of competition-based techniques called genetic algorithms. Genetic algorithms are adaptive search algorithms based on principles derived from the mechanisms of biological evolution. Recent results on the analysis of the implicit parallelism of alternative selection algorithms are summarized, along with an analysis of alternative crossover operators. Applications of these results in practical learning systems for sequential decision problems and for concept classification are also presented. INTRODUCTION One approach to the design of more flexible computer systems is to extract heuristics from existing adaptive systems. We have focused on a class of learning systems that use competition-based procedures, called genetic algorithms (GAs). GAs are ba...

The objective of this research is to develop a system for language learning based on a minimum of pre-wired language-specific functionality, that is compatible with observations of perceptual and language capabilities in the human developmental trajectory. In the proposed system, meaning (in terms of descriptions of events and spatial relations) is extracted from video images based on detection of position, motion, physical contact and their parameters. Mapping of sentence form to meaning is performed by learning grammatical constructions that are retrieved from a construction inventory based on the constellation of closed class items uniquely identifying the target sentence structure. The resulting system displays robust acquisition behavior that reproduces certain observations from developmental studies, with very modest “innate ” language specificity. 1.

for assistance, and Nora Newcombe and Elliott Blass for advice and comments on the manuscript. Above all, I am grateful to Ariel Grace and Kristin Shutts for their unending support and after-hours labor on this project. Draft, 4/20/05. This paper has not yet been peer reviewed. Please do not copy or cite without author's permission. This report considers three prominent claims that boys and men have greater natural aptitude for high-level careers in mathematics and science. According to the first claim, males are more focused on objects and mechanical systems from the beginning of life. According to the second claim, males have a profile of spatial and numerical abilities that predisposes them to greater aptitude in mathematics. According to the third claim, males show greater variability in mathematical aptitude, yielding a preponderance of males at the upper end of the distribution of mathematical talent. Research on cognitive development in human infants and preschool children, and research on cognitive performance by students at all levels, provides evidence against these claims. Mathematical and scientific reasoning develop from a set of biologically based capacities that males and females share. From these capacities, men and women appear to develop equal talent for mathematics and science.

The construction grammar framework provides a rich formalism for the characterization of development in language acquisition, in part because it allows for a graceful transition from idiom-like holophrases, to fully abstract argument constructions. From the perspective of learning complexity, this means that rather than mastering the grammar used by an adult to generate or understand an utterance, the child can exploit more direct correspondences between fixed utterances and their meanings in the form of holophrases, and then progressively apply procedures for generalization in order to develop an abstract argument construction repertoire. The current chapter presents a neuro-computational model that addresses aspects of the construction grammar formalism, and uses this model to explain a possible trajectory from idiom-like holophrases to progressively more abstract argument constructions. The goal of the current research is thus to test a theory of form to meaning mapping that takes substantial input from the proposals of usage based construction grammar (Goldberg 1995, 1998; Croft 2001; Clark 2003, Tomasello 1999, 2003) analogical mapping from sentence to meaning (Fisher 1996), and cue coalitions that specify or identify the mappings (Bates et al. 1982). In particular, the model will provide a functional framework for analysis of the

The current research demonstrates a system inspired by cognitive neuroscience and developmental psychology that learns to construct mappings between the grammatical structure of sentences and the structure of their meaning representations. Sentence to meaning mappings are learned and stored as grammatical constructions. These are stored and retrieved from a construction inventory based on the constellation of closed class items uniquely identifying each construction. These learned mappings allow the system to processes natural language sentences in order to reconstruct complex internal representations of the meanings these sentences describe. The system demonstrates error free performance and systematic generalization for a rich subset of English constructions that includes complex hierarchical grammatical structure, and generalizes systematically to new sentences of the learned construction categories. Further testing demonstrates (1) the capability to accommodate a significantly extended set of constructions, and (2) extension to Japanese, a free word order language that is structurally quite different from English, thus demonstrating the extensibility of the structure mapping model. 1

Abstract not found

SUMMARY—Amid ongoing public speculation about the reasons for sex differences in careers in science and mathematics, we present a consensus statement that is based on the best available scientific evidence. Sex differences in science and math achievement and ability are smaller for the mid-range of the abilities distribution than they are for those with the highest levels of achievement and ability. Males are more variable on most measures of quantitative and visuospatial ability, which necessarily results in more males at both high- and low-ability extremes; the reasons why males are often more variable remain elusive. Successful careers in math and science require many types of cognitive abilities. Females tend to excel in verbal abilities, with large differences between females and males found when assessments include writing

Infants' developing causal expectations for the outcome of a simple tool-use event from ages 8 to 12 months were investigated. Causal expectations were studied by comparing infants' developing tool-use actions (i.e., as tool-use agents) with their developing perceptual reactions (i.e., as tool-use observers) to possible and impossible tool-use events. In Experiment 1, tool-use actions were studied by presenting infants, ages 8 and 12 months, with tool-use object-retrieval problems. In Experiment 2, a second age-matched sample of infants watched a comparable series of possible and impossible tool-use events in which a tool was used to retrieve a goalobject. Two core related findings were made. First, infants' causal action and causal perception develop in parallel. In both action and perception, supporting tool-use develops before surrounding tool-use. Second, infants' tool-use action develops before their causal perception of comparable tool-use events. The findings support the constru...

to talk about events from narrated video in a construction grammar framework

Preferential looking experiments investigated 5- and LX-month-old infants’ perception and understanding of the motions of a shadow that appeared to be cast by a ball upon a box. When all the surfaces within the display were sta-tionary, infants looked reliably longer when the shadow moved than when the shadow was stationary, indicating that they detected the shadow and its motion. In further experiments, however, infants ’ looking was not consistent with a sensitivity to the shadow’s natural motion: They looked longer at nat-ural events in which the shadow moved with the ball or remained at rest under the moving box than at unnatural events in which the shadow moved with the box or remained at rest under the moving ball. These findings suggest that infants overextend to shadows a principle that applies to material objects: Objects move together if and only if they are in contact. In a final experiment, infants were habituated to a moving shadow that repeatedly violated one aspect of the contact principle. In a subsequent test they failed to infer that the shadow would violate another aspect of the contact principle. Instead, they appeared to suspend all predictions concerning the behavior of the shadow. Because environments are made visible by illumination, and because most sources of illumination are directional, shadows are among the most common This research was supported by NIH grant R37 HD 23 103 to ESS and by fellowships from NSF and from the Fyssen Foundation to GAV. Experiments 1-3 were submitted in partial fulfilment of an honors thesis by JSR to the Cornell College Scholars Program. We thank Gavin Huntley-Fenner and Susan Carey for valuable comments and criticism, Deborah King for help with many aspects of the research, and Ed Snyder for building the apparatus.