ISSN:0954-0091    

DOI:10.1080/09540099008915659

出版商:Taylor & Francis Group    

年代:1990

数据来源: Taylor

摘要:

Cognitive linguists hypothesize that language is the product of general cognitive abilities. Semantic and functional motivations are sought for grammatical patterns, sentence meaning is viewed as the result of constraint satisfaction, and highly regular linguistic patterns are thought to be mediated by the same processes as irregular patterns. In this paper, recent cognitive linguistics arguments emphasizing the schematicity continuum, the non-autonomy of syntax, and the non-compositionality of semantics are presented and their amenability to connectionist modeling described. Some of the conceptual matches between cognitive linguistics and connectionism are then illustrated by a back-propagation model of the diverse meanings of the preposition over. The pattern set consisted of a distribution of form-meaning pairs that was meant to be evocative of English usage in that the regularities implicit in the distribution spanned the spectrum from rules to partial regularities to exceptions. Under pressure to encode these regularities with limited resources, the network used one hidden layer to recode the inputs into a set of abstract properties. The properties discovered by the network correspond closely to semantic features that linguists have proposed when giving an account of the meaning of over.

ISSN:0954-0091    

DOI:10.1080/09540099008915660

出版商:Taylor & Francis Group    

年代:1990

数据来源: Taylor

摘要:

In this paper we present a connectionist model of movement in government-binding (GB) theory. The model is a collection of regularly connected groups of connectionist units using only two architectures: a two-dimensional map, used to model tree structures, and a one-dimensional row used to model constituent feature assignment. GB theory is a constraint-based theory of syntax. The constraints of GB are modeled by the maps, the rows and the connections between them. All the connections are regular—they depend only on the maps being linked, not on the individual units. The model performs the movements in the transition from d-structure to s-structure mandated by the constraints on sentence form.

ISSN:0954-0091    

DOI:10.1080/09540099008915661

出版商:Taylor & Francis Group    

年代:1990

数据来源: Taylor

摘要:

There has been much interest in the possibility of connectionist models whose representations can be endowed with compositional structure, and a variety of such models have been proposed. These models typically use distributed representations that arise from the functional composition of constituent parts. Functional composition and decomposition alone, however, yield only an implementation of classical symbolic theories. This paper explores the possibility of moving beyond implementation by exploiting holistic structure-sensitive operations on distributed representations. An experiment is performed using Pollack's Recursive Auto-Associative Memory (RAAM). RAAM is used to construct distributed representations of syntactically structured sentences. A feed-forward network is then trained to operate directly on these representations, modeling syntactic transformations of the represented sentences. Successful training and generalization is obtained, demonstrating that the implicit structure present in these representations can be used for a kind of structure-sensitive processing unique to the connectionist domain.

ISSN:0954-0091    

DOI:10.1080/09540099008915662

出版商:Taylor & Francis Group    

年代:1990

数据来源: Taylor

摘要:

The processing of natural language is, at the same time, naturally symbolic and naturally subsymbolic. It is symbolic because ultimately symbols play a critical role. Writing systems, for example, owe their existence to the symbolic nature of language. It is also subsymbolic because of the nature of speech, the fuzziness of concepts, and the high degree of parallelism that is difficult to explain as a purely symbolic phenomenon. Building a processor of natural language, therefore, requires a hybrid approach. This report details a set of experiments which support the claim that natural language can be syntactically processed in a robust manner using a connectionist deterministic parser. The model is trained from patterns derived from a deterministic grammar and tested with grammatical, ungrammatical and lexically ambiguous sentences.

ISSN:0954-0091    

DOI:10.1080/09540099008915663

出版商:Taylor & Francis Group    

年代:1990

数据来源: Taylor

摘要:

The hierarchical feature map system recognizes an input story as an instance of a particular script by classifying it at three levels: scripts, tracks and role bindings. The recognition taxonomy, i.e. the breakdown of each script into the tracks and roles, is extracted automatically and independently for each script from examples of script instantiations in an unsupervised self-organizing process. The process resembles human learning in that the differentiation of the most frequently encountered scripts become gradually the most detailed. The resulting structure is a hierarchical pyramid of feature maps. The hierarchy visualizes the taxonomy and the maps lay out the topology of each level. The number of input lines and the self-organization time are considerably reduced compared to the ordinary single-level feature mapping. The system can recognize incomplete stories and recover the missing events. The taxonomy also serves as memory organization for script-based episodic memory. The maps assign a unique memory location for each script instantiation. The most salient parts of the input data are separated and most resources are concentrated on representing them accurately.

ISSN:0954-0091    

DOI:10.1080/09540099008915664

出版商:Taylor & Francis Group    

年代:1990

数据来源: Taylor

摘要:

This paper describes a connectionist network which uses surface features and minimal semantic information to identify topic entities within the framework of a simplified representation of the nominal expressions in written narrative discourse. This process is seen as an attentional function critical for interpreting subsequent text and understanding the discourse as a whole. The model also addresses pronoun resolution as a function of attentional state. Since topic entity selection depends heavily on sequentially presented semantic, syntactic and contextual information from the preceding narrative, the model develops an internal representation which incorporates the relevant features of past events in addition to useful features of the current input.

ISSN:0954-0091    

DOI:10.1080/09540099008915665

出版商:Taylor & Francis Group    

年代:1990

数据来源: Taylor

摘要:

Over the last 10 years, the assimilation process referred to as vowel harmony has served as a test case for a number of proposals in phonological theory. Current autosegmental approaches successfully capture the intuition that vowel harmony is a dynamic process involving the interaction of a sequence of vowels; still, no theoretical analysis has offered a non-stipulative account of the inconsistent behavior of the so-called ‘transparent’, or disharmonic, segments. This paper proposes a connectionist processing account of the vowel harmony phenomenon, using data from Hungarian. The strength of this account is that it demonstrates that the same general principle of assimilation which underlies the behavior of the 'harmonic' forms accounts as well for the apparently exceptional "transparent' cases, without stipulation. The account proceeds in three steps. After presenting the data and current theoretical analyses, the paper describes the model of sequential processing introduced by Jordan (1986), and motivates this as a model of assimilation processes in phonology. The paper then presents the results of a series of parametric studies that were run with this model, using arbitrary bit patterns as stimuli. These results establish certain conditions on assimilation in a network of this type. Finally, these findings are related to the Hungarian data, where the same conditions are shown to predict the correct pattern of behavior for both harmonic and transparent vowels.

ISSN:0954-0091    

DOI:10.1080/09540099008915666

出版商:Taylor & Francis Group    

年代:1990

数据来源: Taylor

摘要:

How can a nervous system represent for itself the temporal relations of patterns that it knows? In order to label auditory patterns, the nervous system must store early portions in order to identify the whole. Both linguists and engineer-scientists have a similar need to record spoken words. This paper reviews three basic models for handling the information-collection problem that supports pattern recognition, whether by scientists or others. Many of these techniques have been implemented in connectionist networks. In linguistic models for words, there are only ordered symbols, i.e. either phonemic segments or words. In engineering and speech science, time windows are built that store the entire signal and allow parametric description of time. But such windows are not plausible for nervous systems. A third alternative is a memory in the form of a dynamic system. These models are driven through a trajectory in state space by the input signals. Thus, the recognition process for familiar patterns produces a distinct trajectory through state space for each learned pattern. Among the advantages of such a system are that (1) it tends to recognize patterns despite changes in the rate of presentation, and (2) the system can be run continuously yet will respond as quickly as possible at appropriate times. Evidence is reviewed about human auditory memory for complex tone sequences. The data suggest that human auditory memory exhibits many similarities to the dynamic model.

ISSN:0954-0091    

DOI:10.1080/09540099008915667

出版商:Taylor & Francis Group    

年代:1990

数据来源: Taylor