Words with Consistent Diachronic Usage Patterns are Learned Earlier: A Computational Analysis Using Temporally Aligned Word Embeddings

In this study, we use temporally aligned word embeddings and a large diachronic corpus of English to quantify language change in a data-driven, scalable way, which is grounded in language use. We show a unique and reliable relation between measures of language change and age of acquisition (AoA) while controlling for frequency, contextual diversity, concreteness, length, dominant part of speech, orthographic neighborhood density, and diachronic frequency variation. We analyze measures of language change tackling both the change in lexical representations and the change in the relation between lexical representations and the words with the most similar usage patterns, showing that they capture different aspects of language change. Our results show a unique relation between language change and AoA, which is stronger when considering neighborhood-level measures of language change: Words with more coherent diachronic usage patterns tend to be acquired earlier. The results support theories positing a link between ontogenetic and ethnogenetic processes in language.     

A Computational and Empirical Investigation of Graphemes in Reading

It is often assumed that graphemes are a crucial level of orthographic representation above letters. Current connectionist models of reading, however, do not address how the mapping from letters to graphemes is learned. One major challenge for computational modeling is therefore developing a model that learns this mapping and can assign the graphemes to linguistically meaningful categories such as the onset, vowel, and coda of a syllable. Here, we present a model that learns to do this in English for strings of any letter length and any number of syllables. The model is evaluated on error rates and further validated on the results of a behavioral experiment designed to examine ambiguities in the processing of graphemes. The results show that the model (a) chooses graphemes from letter strings with a high level of accuracy, even when trained on only a small portion of the English lexicon; (b) chooses a similar set of graphemes as people do in situations where different graphemes can potentially be selected; (c) predicts orthographic effects on segmentation which are found in human data; and (d) can be readily integrated into a full‐blown model of multi‐syllabic reading aloud such as CDP++ (Perry, Ziegler, & Zorzi, 2010). Altogether, these results suggest that the model provides a plausible hypothesis for the kind of computations that underlie the use of graphemes in skilled reading.     

Direct Versus Indirect Causation as a Semantic Linguistic Universal: Using a Computational Model of English,Hebrew, Hindi,Japanese, and K'iche’ Mayan to Predict Grammaticality Judgments in Balinese

The aim of this study was to test the claim that languages universally employ morphosyntactic marking to differentiate events of more- versus less-direct causation, preferring to mark them with less- and more- overt marking, respectively (e.g., Somebody broke the window vs. Somebody MADE the window break; *Somebody cried the boy vs. Somebody MADE the boy cry). To this end, we investigated whether a recent computational model which learns to predict speakers’ by-verb relative preference for the two causatives in English, Hebrew, Hindi, Japanese, and K'iche’ Mayan is able to generalize to a sixth language on which it has never been trained: Balinese. Judgments of the relative acceptability of the less- and more-transparent causative forms of 60 verbs were collected from 48 native-speaking Balinese adults. The composite crosslinguistic computational model was able to predict these judgments, not only for verbs that it had seen, but also––in a split-half validation test––to verbs that it had never seen in any language. A “random-semantics” model showed only a relatively small decrement in performance with seen verbs, whose behavior can be learned on a verb-by-verb basis, but achieved zero correlation with human judgments when generalizing to unseen verbs. Together, these findings suggest that Balinese conceptualizes directness of causation in a similar way to these unrelated languages, and therefore constitute support for the view that the distinction between more- versus less-distinct causation constitutes a morphosyntactic universal.     

Perceiving versus inferring movements to understand dynamic events: The influence of content complexity

ObjectivesWhat type of visual presentation is best in helping learners to understand the functioning of a dynamic system and under what conditions? This study investigated the effect of content complexity on perceived cognitive load and performance resulting from studying depicted movements of team play either in an explicit manner (animation) or via arrow symbols (static diagram).DesignA 2 (treatment: diagram vs. animation) × 2 (content complexity: low vs. high) between subjects design was adopted in the experiment.MethodsForty-eight university students were randomly assigned to the four study conditions and required to perform a reconstruction test and rate their perceived cognitive load following the study phase.ResultsData analyses revealed that for low-complexity content, participants exposed to the animation treatment learned more efficiently – based on the combination of learning and cognitive load scores – than those exposed to the diagram treatment. On the other hand, for high-complexity content, participants exposed to the diagram treatment learned more efficiently than those exposed to the animation treatment.ConclusionThe findings stress the importance of considering the task complexity factor when designing and presenting instructional materials to learners.     

MetrIntMeas a novel metric for measuring the intelligence of a swarm of cooperating agents

We propose a novel metric called MetrIntMeas (Metric for the Intelligence Measuring) for an accurate and robust measurement of the difficult problem-solving intelligence of a swarm system. The metric allows the classification if a swarm system belongs to the same class with the systems which have a specific reference intelligence value. For proving the efficiency of the proposed metric we realized a case study on a swarm system specialized in solving a NP-hard problem. As an application of the proposed metric, we present the measurement of the swarm systems’ evolution in intelligence. We gave a new definition to the intelligent evolving systems. The evolution of intelligent systems can be verified using the proposed MetrIntMeas metric.