The literature contains a disconnect between accounts of how humans learn lexical semantic representations for words. Theories generally propose that lexical semantics are learned either through perceptual experience or through exposure to regularities in language. We propose here a model to integrate these two information sources. Specifically, the model uses the global structure of memory to exploit the redundancy between language and perception in order to generate inferred perceptual representations for words with which the model has no perceptual experience. We (...) test the model on a variety of different datasets from grounded cognition experiments and demonstrate that this diverse set of results can be explained as perceptual simulation (cf. Barsalou, Simmons, Barbey, & Wilson, 2003) within a global memory model. (shrink)

Verbal narratives provide incomplete information and can be very long, yet readers and hearers often effortlessly fill in the gaps and make connections across long stretches of text, sometimes even finding this immersive. How is this done? In the last few decades, event-indexing situation modeling and complementary accounts of narrative emotion have suggested answers. Despite this progress, comparisons between real-life perception and narrative experience might underplay the way narrative processing modifies our world model, as well as the role of the (...) emotions that do not relate to characters. I reframe narrative experience in predictive processing and neural networks, capturing continuity between fiction, perception, and states like dreaming and imagination, enabled by the flexible instantiation of concepts. In this framework, narrative experience is more clearly revealed as a creative experience that can share some of the phenomenology of dreams. (shrink)

The computations involved in statistical learning have long been debated. Here, we build on work suggesting that a basic memory process, chunking, may account for the processing of statistical regularities into larger units. Drawing on methods from the memory literature, we developed a novel paradigm to test statistical learning by leveraging a robust phenomenon observed in serial recall tasks: that short‐term memory is fundamentally shaped by long‐term distributional learning. In the statistically induced chunking recall (SICR) task, participants are exposed to (...) an artificial language, using a standard statistical learning exposure phase. Afterward, they recall strings of syllables that either follow the statistics of the artificial language or comprise the same syllables presented in a random order. We hypothesized that if individuals had chunked the artificial language into word‐like units, then the statistically structured items would be more accurately recalled relative to the random controls. Our results demonstrate that SICR effectively captures learning in both the auditory and visual modalities, with participants displaying significantly improved recall of the statistically structured items, and even recall specific trigram chunks from the input. SICR also exhibits greater test–retest reliability in the auditory modality and sensitivity to individual differences in both modalities than the standard two‐alternative forced‐choice task. These results thereby provide key empirical support to the chunking account of statistical learning and contribute a valuable new tool to the literature. (shrink)

Natural language is full of patterns that appear to fit with general linguistic rules but are ungrammatical. There has been much debate over how children acquire these “linguistic restrictions,” and whether innate language knowledge is needed. Recently, it has been shown that restrictions in language can be learned asymptotically via probabilistic inference using the minimum description length (MDL) principle. Here, we extend the MDL approach to give a simple and practical methodology for estimating how much linguistic data are required to (...) learn a particular linguistic restriction. Our method provides a new research tool, allowing arguments about natural language learnability to be made explicit and quantified for the first time. We apply this method to a range of classic puzzles in language acquisition. We find some linguistic rules appear easily statistically learnable from language experience only, whereas others appear to require additional learning mechanisms (e.g., additional cues or innate constraints). (shrink)

Children learn their native language by exposure to their linguistic and communicative environment, but apparently without requiring that their mistakes be corrected. Such learning from “positive evidence” has been viewed as raising “logical” problems for language acquisition. In particular, without correction, how is the child to recover from conjecturing an over-general grammar, which will be consistent with any sentence that the child hears? There have been many proposals concerning how this “logical problem” can be dissolved. In this study, we review (...) recent formal results showing that the learner has sufficient data to learn successfully from positive evidence, if it favors the simplest encoding of the linguistic input. Results include the learnability of linguistic prediction, grammaticality judgments, language production, and form-meaning mappings. The simplicity approach can also be “scaled down” to analyze the learnability of specific linguistic constructions, and it is amenable to empirical testing as a framework for describing human language acquisition. (shrink)

Computational models of semantic memory exploit information about co-occurrences of words in naturally occurring text to extract information about the meaning of the words that are present in the language. Such models implicitly specify a representation of temporal context. Depending on the model, words are said to have occurred in the same context if they are presented within a moving window, within the same sentence, or within the same document. The temporal context model (TCM), which specifies a particular definition of (...) temporal context, has proved useful in the study of episodic memory. The predictive temporal context model (pTCM) uses the same definition of temporal context to generate semantic memory representations. Taken together pTCM and TCM may prove to be part of a general model of declarative memory. (shrink)

It is possible to learn multiple layers of non-linear features by backpropagating error derivatives through a feedforward neural network. This is a very effective learning procedure when there is a huge amount of labeled training data, but for many learning tasks very few labeled examples are available. In an effort to overcome the need for labeled data, several different generative models were developed that learned interesting features by modeling the higher order statistical structure of a set of input vectors. One (...) of these generative models, the restricted Boltzmann machine (RBM), has no connections between its hidden units and this makes perceptual inference and learning much simpler. More significantly, after a layer of hidden features has been learned, the activities of these features can be used as training data for another RBM. By applying this idea recursively, it is possible to learn a deep hierarchy of progressively more complicated features without requiring any labeled data. This deep hierarchy can then be treated as a feedforward neural network which can be discriminatively fine-tuned using backpropagation. Using a stack of RBMs to initialize the weights of a feedforward neural network allows backpropagation to work effectively in much deeper networks and it leads to much better generalization. A stack of RBMs can also be used to initialize a deep Boltzmann machine that has many hidden layers. Combining this initialization method with a new method for fine-tuning the weights finally leads to the first efficient way of training Boltzmann machines with many hidden layers and millions of weights. (shrink)

Much of traditional AI exemplifies the explicit representation paradigm, and during the late 1980''s a heated debate arose between the classical and connectionist camps as to whether beliefs and rules receive an explicit or implicit representation in human cognition. In a recent paper, Kirsh (1990) questions the coherence of the fundamental distinction underlying this debate. He argues that our basic intuitions concerning explicit and implicit representations are not only confused but inconsistent. Ultimately, Kirsh proposes a new formulation of the distinction, (...) based upon the criterion ofconstant time processing.The present paper examines Kirsh''s claims. It is argued that Kirsh fails to demonstrate that our usage of explicit and implicit is seriously confused or inconsistent. Furthermore, it is argued that Kirsh''s new formulation of the explicit-implicit distinction is excessively stringent, in that it banishes virtually all sentences of natural language from the realm of explicit representation. By contrast, the present paper proposes definitions for explicit and implicit which preserve most of our strong intuitions concerning straightforward uses of these terms. It is also argued that the distinction delineated here sustains the meaningfulness of the abovementioned debate between classicists and connectionists. (shrink)

In their provocative 1988 paper, Fodor and Pylyshyn issued a formidable challenge to connectionists, i.e. to provide a non‐classical explanation of the empirical phenomenon of systematicity in cognitive agents. Since the appearance of F&P's challenge, a number of connectionist systems have emerged which prima facie meet this challenge. However, Fodor and McLaughlin (1990) advance an argument, based upon a general principle of nomological necessity, to show that one of these systems (Smolensky's) could not satisfy the Fodor‐Pylyshyn challenge. Yet, if Fodor (...) and McLaughlin's analysis is correct, it is doubtful whether any existing connectionist system would fare better than Smolensky's. In the view of Fodor and McLaughlin, humans and classical architectures display systematicity as a matter of nomological necessity (necessity by virtue of natural law), but connectionist architectures do not. However, I argue that the Fodor‐Pylyshyn‐McLaughlin appeal to nomological necessity is untenable. There is a sense in which neither classical nor connectionist architectures possess nomological (or‘nomic’) necessity. However, the sense in which classical architectures do possess nomic necessity applies equally well to at least some connectionist architectures. Representational constituents can have causal efficacy within both classical and connectionist architectures. (shrink)

In their provocative 1988 paper, Fodor and Pylyshyn issued a formidable challenge to connectionists, i.e. to provide a non‐classical explanation of the empirical phenomenon of systematicity in cognitive agents. Since the appearance of F&P's challenge, a number of connectionist systems have emerged which prima facie meet this challenge. However, Fodor and McLaughlin (1990) advance an argument, based upon a general principle of nomological necessity, to show that one of these systems (Smolensky's) could not satisfy the Fodor‐Pylyshyn challenge. Yet, if Fodor (...) and McLaughlin's analysis is correct, it is doubtful whether any existing connectionist system would fare better than Smolensky's. In the view of Fodor and McLaughlin, humans and classical architectures display systematicity as a matter of nomological necessity (necessity by virtue of natural law), but connectionist architectures do not. However, I argue that the Fodor‐Pylyshyn‐McLaughlin appeal to nomological necessity is untenable. There is a sense in which neither classical nor connectionist architectures possess nomological (or‘nomic’) necessity. However, the sense in which classical architectures do possess nomic necessity applies equally well to at least some connectionist architectures. Representational constituents can have causal efficacy within both classical and connectionist architectures. (shrink)

A large body of psycholinguistic research demonstrates that both language processing and language acquisition are sensitive to the distributions of linguistic constructions in usage. Here we investigate how statistical distributions at different linguistic levels – morphological and lexical, and phrasal – contribute to the ease with which morphosyntax is processed and produced by second language learners. We analyze Chinese ESL learners’ knowledge of four English inflectional morphemes: -ed, -ing, and third-person -s on verbs, and plural -s on nouns. In Elicited (...) Imitation Tasks, participants listened to length- and difficulty-matched sentences each containing one target morpheme and typed the whole sentence as accurately as they could after a short delay. Experiment 1 investigated lexical and morphemic levels, testing the hypotheses that a morpheme is expected to be more easily processed when it is highly available, and highly reliable. Thirty sentences were made for each morpheme, divided into three Availability-Reliability Distribution groups on the basis of corpus analysis in the Corpus of Contemporary American English : 10 target words high in availability, 10 high in reliability, and 10 low in both reliability and availability. Responses were scored on whether the target morpheme was accurately reproduced given the provision of the correct lemma. A generalized linear mixed-effects logit model revealed fixed effects of morpheme type, availability, and reliability on the accuracy of morpheme provision. There were no effects of lemma frequency. Experiment 2 successfully replicated these results and extended the investigation to explore phrasal formulaicity by manipulating the frequency of the four-word strings in which the morpheme was embedded. GLMMs replicated the effects of word-form availability and reliability and additionally revealed independent phrase-superiority effects where morphemes were better reproduced in contexts of higher string-frequency. Taken together, these findings demonstrate that morpheme acquisition reflects the distributional properties of learners’ experience and the mappings therein between lexis, morphology, phraseology, and semantics. These conclusions support an emergentist view of the statistical symbolic learning of morphology where language acquisition involves the satisfaction of competing constraints across multiple grain-sizes of units. (shrink)

‘Particularism’ and ‘generalism’ refer to families of positions in the philosophy of moral reasoning, with the former playing down the importance of principles, rules or standards, and the latter stressing their importance. Part of the debate has taken an empirical turn, and this turn has implications for AI research and the philosophy of cognitive modeling. In this paper, Jonathan Dancy’s approach to particularism (arguably one of the best known and most radical approaches) is questioned both on logical and empirical grounds. (...) Doubts are raised over whether Dancy’s brand of particularism can adequately explain the graded nature of similarity assessments in analogical arguments. Also, simple recurrent neural network models of moral case classification are presented and discussed. This is done to raise concerns about Dancy’s suggestion that neural networks can help us to understand how we could classify situations in a way that is compatible with his particularism. Throughout, the idea of a surveyable standard—one with restricted length and complexity—plays a key role. Analogical arguments are taken to involve multidimensional similarity assessments, and surveyable contributory standards are taken to be attempts to articulate the dimensions of similarity that may exist between cases. This work will be of relevance both to those who have interests in computationally modeling human moral cognition and to those who are interested in how such models may or may not improve our philosophical understanding of such cognition. (shrink)

The human brain is autonomously active, being characterized by a self-sustained neural activity which would be present even in the absence of external sensory stimuli. Here we study the interrelation between the self-sustained activity in autonomously active recurrent neural nets and external sensory stimuli. There is no a priori semantical relation between the influx of external stimuli and the patterns generated internally by the autonomous and ongoing brain dynamics. The question then arises when and how are semantic correlations between internal (...) and external dynamical processes learned and built up? We study this problem within the paradigm of transient state dynamics for the neural activity in recurrent neural nets, i.e. for an autonomous neural activity characterized by an infinite time-series of transiently stable attractor states. We propose that external stimuli will be relevant during the sensitive periods, viz the transition period between one transient state and the subsequent semi-stable attractor. A diffusive learning signal is generated unsupervised whenever the stimulus influences the internal dynamics qualitatively. For testing we have presented to the model system stimuli corresponding to the bars and stripes problem. We found that the system performs a non-linear independent component analysis on its own, being continuously and autonomously active. This emergent cognitive capability results here from a general principle for the neural dynamics, the competition between neural ensembles. (shrink)

Many of the problems studied in cognitive science are inductive problems, requiring people to evaluate hypotheses in the light of data. The key to solving these problems successfully is having the right inductive biases—assumptions about the world that make it possible to choose between hypotheses that are equally consistent with the observed data. This article explores a novel experimental method for identifying the biases that guide human inductive inferences. The idea behind this method is simple: This article uses the responses (...) produced by a participant on one trial to generate the stimuli that either they or another participant will see on the next. A formal analysis of this “iterated learning” procedure, based on the assumption that the learners are Bayesian agents, predicts that it should reveal the inductive biases of these learners, as expressed in a prior probability distribution over hypotheses. This article presents a series of experiments using stimuli based on a well-studied set of category structures, demonstrating that iterated learning can be used to reveal the inductive biases of human learners. (shrink)

Theories of sentence production that involve a convergence of activation from conceptual‐semantic and syntactic‐sequential units inspired a connectionist model that was trained to produce simple sentences. The model used a learning algorithm that resulted in a sharing of responsibility (or “division of labor”) between syntactic and semantic inputs for lexical activation according to their predictive power. Semantically rich, or “heavy”, verbs in the model came to rely on semantic cues more than on syntactic cues, whereas semantically impoverished, or “light”, verbs (...) relied more on syntactic cues. When the syntactic and semantic inputs were lesioned, the model exhibited patterns of production characteristic of agrammatic and anomic aphasic patients, respectively. Anomic models tended to lose the ability to retrieve heavy verbs, whereas agrammatic models were more impaired in retrieving light verbs. These results obtained in both sentence production and single‐word naming simulations. Moreover, simulated agrammatic lexical retrieval was more impaired overall in sentences than in single‐word tasks, in agreement with the literature. The results provide a demonstration of the division‐of‐labor principle, as well as general support for the claim that connectionist learning principles can contribute to the understanding of non‐transparent neuropsychological dissociations. (shrink)

Saffran, Newport, and Aslin (1996a) found that human infants are sensitive to statistical regularities corresponding to lexical units when hearing an artificial spoken language. Two sorts of segmentation strategies have been proposed to account for this early word‐segmentation ability: bracketing strategies, in which infants are assumed to insert boundaries into continuous speech, and clustering strategies, in which infants are assumed to group certain speech sequences together into units (Swingley, 2005). In the present study, we test the predictions of two computational (...) models instantiating each of these strategies i.e., Serial Recurrent Networks: Elman, 1990; and Parser: Perruchet & Vinter, 1998 in an experiment where we compare the lexical and sublexical recognition performance of adults after hearing 2 or 10 min of an artificial spoken language. The results are consistent with Parser’s predictions and the clustering approach, showing that performance on words is better than performance on part‐words only after 10 min. This result suggests that word segmentation abilities are not merely due to stronger associations between sublexical units but to the emergence of stronger lexical representations during the development of speech perception processes. (shrink)

Many natural and technological systems are complex, with organizational structures that exhibit characteristic patterns but defy concise description. One effective approach to analyzing such systems is in terms of repeated topological motifs. Here, we extend the motif concept to characterize the dynamic behavior of complex systems by introducing developmental motifs, which capture patterns of system growth. As a proof of concept, we use developmental motifs to analyze the developmental cell lineage of the nematode Caenorhabditis elegans, revealing a new perspective on (...) its complex structure. We use a family of computational models to explore how biases arising from the dynamics of the developmental gene network, as well as spatial and temporal constraints acting on development, contribute to this complex organization. © 2010 Wiley Periodicals, Inc. Complexity 16: 48‐57, 2011. (shrink)

The present paper intends to draw the conception of language implied in the technique of word embeddings that supported the recent development of deep neural network models in computational linguistics. After a preliminary presentation of the basic functioning of elementary artificial neural networks, we introduce the motivations and capabilities of word embeddings through one of its pioneering models, word2vec. To assess the remarkable results of the latter, we inspect the nature of its underlying mechanisms, which have been characterized as the (...) implicit factorization of a word-context matrix. We then discuss the ordinary association of the “distributional hypothesis” with a “use theory of meaning,” often justifying the theoretical basis of word embeddings, and contrast them to the theory of meaning stemming from those mechanisms through the lens of matrix models (such as vector space models and distributional semantic models). Finally, we trace back the principles of their possible consistency through Harris’s original distributionalism up to the structuralist conception of language of Saussure and Hjelmslev. Other than giving access to the technical literature and state of the art in the field of natural language processing to non-specialist readers, the paper seeks to reveal the conceptual and philosophical stakes involved in the recent application of new neural network techniques to the computational treatment of language. (shrink)