Anchored Packed Trees (APTs) are a novel approach to distributional semantics that takes distributional composition to be a process of lexeme contextualisation. A lexeme's meaning, characterised as knowledge concerning co-occurrences involving that lexeme, is represented with a higher-order dependency-typed structure (the APT) where paths associated with higher-order dependencies connect vertices associated with weighted lexeme multisets. The central innovation in the compositional theory is that the APT's type structure enables the precise alignment of the semantic representation of each of the lexemes being composed. Like other count-based distributional spaces, however, Anchored Packed Trees are prone to considerable data sparsity, caused by not observing all plausible co-occurrences in the given data. This problem is amplified for models like APTs, that take the grammatical type of a co-occurrence into account. This results in a very sparse distributional space, requiring a mechanism for inferring missing knowledge. Most methods face this challenge in ways that render the resulting word representations uninterpretable, with the consequence that distributional composition becomes difficult to model and reason about. In this thesis, I will present a practical evaluation of the Apt theory, including a large-scale hyperparameter sensitivity study and a characterisation of the distributional space that APTs give rise to. Based on the empirical analysis, the impact of the problem of data sparsity is investigated. In order to address the data sparsity challenge and retain the interpretability of the model, I explore an alternative algorithm — distributional inference — for improving elementary representations. The algorithm involves explicitly inferring unobserved co-occurrence events by leveraging the distributional neighbourhood of the semantic space. I then leverage the rich type structure in APTs and propose a generalisation of the distributional inference algorithm. I empirically show that distributional inference improves elementary word representations and is especially beneficial when combined with an intersective composition function, which is due to the complementary nature of inference and composition. Lastly, I qualitatively analyse the proposed algorithms in order to characterise the knowledge that they are able to infer, as well as their impact on the distributional APT space.

This dissertation proposes a general methodological framework for the application of computational text analysis to the study of long duration material processes of transformation, beyond their traditional application to the study of discourse and rhetorical action. Over a thin theory of the linguistic nature of social facts, the proposed methodology revolves around the compilation of term co-occurrence matrices and their projection into different representations of an hypothetical semantic space. These representations offer solutions to two problems inherent to social scientific research: that of "mapping" features in a given representation to theoretical entities and that of "alignment" of the features seen in models built from different sources in order to enable their comparison. The data requirements of the exercise are discussed through the introduction of the notion of a "narrative horizon", the extent to which a given source incorporates a narrative account in its rendering of the context that produces it. Useful primary data will consist of text with short narrative horizons, such that the ideal source will correspond to a continuous archive of institutional, ideally bureaucratic text produced as mere documentation of a definite population of more or less stable and comparable social facts across a couple of centuries. Such a primary source is available in the Proceedings of the Old Bailey (POB), a collection of transcriptions of 197,752 criminal trials seen by the Old Bailey and the Central Criminal Court of London and Middlesex between 1674 and 1913 that includes verbatim transcriptions of witness testimony. The POB is used to demonstrate the proposed framework, starting with the analysis of the evolution of an historical corpus to illustrate the procedure by which provenance data is used to construct longitudinal and cross-sectional comparisons of different corpus segments. The co-occurrence matrices obtained from the POB corpus are used to demonstrate two different projections: semantic networks that model different notions of similarity between the terms in a corpus' lexicon as an adjacency matrix describing a graph and semantic vector spaces that approximate a lower-dimensional representation of an hypothetical semantic space from its empirical effects on the co-occurrence matrix. Semantic networks are presented as discrete mathematical objects that offer a solution to the mapping problem through operation that allow for the construction of sets of terms over which an order can be induced using any measure of significance of the strength of association between a term set and its elements. Alignment can then be solved through different similarity measures computed over the intersection and union of the sets under comparison. Semantic vector spaces are presented as continuous mathematical objects that offer a solution to the mapping problem in the linear structures contained in them. This include, in all cases, a meaningful metric that makes it possible to define neighbourhoods and regions in the semantic space and, in some cases, a meaningful orientation that makes it possible to trace dimensions across them. Alignment can then proceed endogenously in the case of oriented vector spaces for relative comparisons, or through the construction of common basis sets for non-oriented semantic spaces for absolute comparisons. The dissertation concludes with the proposition of a general research program for the systematic compilation of text distributional patterns in order to facilitate a much needed process of calibration required by the techniques discussed in the previous chapters. Two specific avenues for further research are identified. First, the development of incremental methods of projection that allow a semantic model to be updated as new observations come along, an area that has received considerable attention from the field of electronic finance and the pervasive use of Gentleman's algorithm for matrix factorisation. Second, the development of additively decomposable models that may be combined or disaggregated to obtain a similar result to the one that would have been obtained had the model being computed from the union or difference of their inputs. This is established to be dependent on whether the functions that actualise a given model are associative under addition or not.

Natural Language Processing systems crucially depend on the availability of lexical and conceptual knowledge representations. They need to be able to disambiguate word senses and detect synonyms. In order to draw inferences, they require access to hierarchical relations between concepts (dog isAn animal) as well as non-hierarchical ones (gasoline fuels car). Knowledge resources such as lexical databases, semantic networks and ontologies explicitly encode such conceptual knowledge. However, traditionally, these have been manually created, which is expensive and time consuming for large re- sources, and cannot provide adequate coverage in specialised domains. In order to alleviate this acquisition bottleneck, statistical methods have been created to acquire lexical and conceptual knowledge automatically from text. In particular, unsupervised techniques have the advantage that they can be easily adapted to any domain, given some corpus on the topic. However, due to sparseness issues, they often require very large corpora to achieve high quality results. The spectrum of resources and statistical methods has a crucial gap in situations when manually cre- ated resources do not provide the necessary coverage and only limited corpora are available. This is the case for real-world domain applications such as an NLP system for processing technical information based on a limited amount of company documentation. We provide a large-scale demonstration that this gap can be filled through the use of automatically generated networks. The corpus is automatically transformed into a network representing the terms or concepts which occur in the text and their relations, based entirely on linguistic tools. The net- works structurally lie in between the unstructured corpus and the highly structured manually created resources. We show that they can be useful in situations for which neither existing approach is ap- plicable. In contrast to manually created resources, our networks can be generated quickly and on demand. Conversely, they make it possible to achieve higher quality representations from less text than corpus-based methods, relieving the requirement of very large scale corpora. We devise scaleable frameworks for building networks from plain text and Wikipedia with varying levels of expressiveness. This work creates concrete networks from the entire British National Corpus covering 1.2m terms and 21m relations and a Wikipedia network covering 2.7m concepts. We develop a network-based semantic space model and evaluate it on the task of measuring semantic relatedness. In addition, noun compound paraphrasing is tackled to demonstrate the quality of the indirect paths in the network for concept relation description. On both evaluations we achieve results competitive to the state of the art. In particular, our network-based methods outperform corpus-based methods, demonstrating the gain created by leveraging the network structure.

This paper is about an implemented natural language interface that translates from English into semantic net relations and from semantic net relations back into English. The parser and companion generator were implemented for two reasons: (a) to enable experimental work in support of a theory of learning by analogy; (b) to demonstrate the viability of a theory of parsing and generation built on commutative transformations. The learning theory was shaped to a great degree by experiments that would have been extraordinarily tedious to perform without the English interface with which the experimental data base was prepared, revise, and revised again. Inasmuch as current work on the learning theory is moving toward a tenfold increase in data-base size, the English interface is moving from a facilitating role to an enabling one. The parsing and generation theory has two particularly important features: (a) the same grammar is used for both parsing and generation; (b) the transformations of the grammar are commutative. The language generation procedure converts a semantic network fragment into kernel frames, chooses the set of transformations that should be performed upon each frame, executes the specified transformations, combines the altered kernels into a sentence, performs a pronominalization process, and finally produces the appropriate English word string. Parsing is essentially the reverse of generation. The first step in the parsing process is splitting a given sentence into a set of kernel clauses along with a description of how those clauses hierarchically related to each other. The clauses are hierarchically related to each other. The clauses are used to produce a matrix embedded kernel frames, which in turn supply arguments to relation-creating functions. The evaluation of the relation-creating functions results in the construction of the semantic net fragments.

Omega is a description system for knowledge embedding which incorporates some of the attractive modes of expression in common sense reasoning such as descriptions, inheritance, quantification, negation, attributions and multiple viewpoints. A formalization of Omega is developed as a framework for investigations on the foundations of knowledge representation. As a logic, Omega achieves the goal of an intuitively sound and consistent theory of classes which permits unrestricted abstraction within a powerful logic system. Description abstraction is the construct provided in Omega corresponding to set abstraction. Attributions and inheritance are the basic mechanisms for knowledge structuring. To achieve flexibility and incrementality, the language allows descriptions with an arbitrary number of attributions, rather than predicates with a fixed number of arguments as in predicate logic. This requires a peculiar interpretation for instance descriptions, which in turn provides insights into the use and meaning of several kinds of attributions. The formal treatment consists in presenting semantic models for Omega, deriving an axiomatization and establishing the consistency and completeness of the logic.

Semantic Routed Networks provide a superior infrastructure for complex search engines. In a Semantic Routed Network (SRN), the routers are the critical component and they perform semantic comparison as their key computation. As the amount of information available on the Internet grows, the speed and efficiency with which information can be retrieved to the user becomes important. Most current search engines scale to meet the growing demand by deploying large data centers with general purpose computers that consume many megawatts of power. Reducing the power consumption of these data centers while providing better performance, will help reduce the costs of operation significantly. Performing operations in parallel is a key optimization step for better performance on general purpose CPUs. Current techniques for parallelization include architectures that are multi-core and have multiple thread handling capabilities. These coarse grained approaches have considerable resource management overhead and provide only sub-linear speedup. This dissertation proposes techniques towards a highly parallel, power efficient architecture that performs semantic comparisons as its core activity. Hardware-centric parallel algorithms have been developed to populate the required data structures followed by computation of semantic similarity. The performance of the proposed design is further enhanced using a pipelined architecture. The proposed algorithms were also implemented on two contemporary platforms such as the Nvidia CUDA and an FPGA for performance comparison. In order to validate the designs, a semantic benchmark was also been created. It has been shown that a dedicated semantic comparator delivers significantly better performance compared to other platforms. Results show that the proposed hardware semantic comparison architecture delivers a speedup performance of up to 10^5 while reducing power consumption by 80% compared to traditional computing platforms. Future research directions including better power optimization, architecting the complete semantic router and using the semantic benchmark for SRN research are also discussed.

Thesis (Ph. D.)--University of California, San Diego, 1997. / Vita. Includes bibliographical references (leaves 194-202).

Λαμβάνοντας υπόψη ότι η χρήση των υπολογιστών μεταφέρθηκε από τους ερευνητικούς οργανισμούς στις επιχειρήσεις, διαπιστώνουμε ότι αποτελεί πλέον αναπόσπαστο επιχειρηματικό εργαλείο. Φτάνοντας στις αρχές της δεκαετίας του ‘90, μεγάλοι επιχειρηματικοί και κρατικοί φορείς διέθεταν τεράστιες ποσότητες δεδομένων που θα έπρεπε να εκμεταλλευθούν. Τα παραπάνω σε συνδυασμό με την εδραιωμένη πλέον αντίληψη ότι η πληροφορία αποτελεί το πιο πολύτιμο αγαθό, οδήγησαν στην ανάγκη για εφαρμογές ανάλυσης και επεξεργασίας μεγάλου όγκου δεδομένων. Την λύση δίνουν οι τεχνολογίες των Αποθηκών Δεδομένων (Data warehouses) και της αναλυτικής Επεξεργασίας Δεδομένων (OLAP). Τον τελευταίο καιρό διεξάγεται σημαντική έρευνα σε οντολογίες και σημασιολογικά δίκτυα, όπου πλέον η πληροφορία περιγράφεται εννοιολογικά για να είναι ευκολότερη η ανάκτηση της, η χρησιμοποίηση της και η σύγκριση της, δημιουργείται η ανάγκη εύρεσης ενός νέου συνδυαστικού τρόπου αναπαράστασης της γνώσης. Στην παρούσα διπλωματική εργασία ο αναγνώστης εισάγεται στα σημασιολογικά δίκτυα. Εκεί παρουσιάζονται αναλυτικά τα είδη των σημασιολογικών δικτύων και ιδιαίτερα ορισμένα που χρησιμοποιούνται για την αντιμετώπιση εξαιρέσεων. Δίνετε εισαγωγική περιγραφή της θεωρίας που αναπτύσσεται στις βάσεις δεδομένων με ιδιαίτερη έμφαση στις πολυδιάστατες βάσεις και τις αποθήκες δεδομένων. Παρουσιάζονται αναλυτικά τα OLAP εργαλεία καθώς και οι βασικές έννοιες που τα αποτελούν. Τέλος, γίνεται η παρουσίαση μιας νέας τεχνικής που βασίζεται στις μεθόδους αντιμετώπισης εξαιρέσεων σε ένα σημασιολογικό δίκτυο με σκοπό να επεκτείνει ένα OLAP μοντέλο και να το καταστήσει ικανό να αντιμετωπίσει εξαιρέσεις. Η προτεινόμενη επέκταση επιτρέπει την ύπαρξη εξαιρέσεων μεταξύ των τιμών των διαστάσεων ενός υπερκύβου δεδομένων, οι οποίες παίζουν σημαντικό ρόλο στην σωστή εξαγωγή συμπερασμάτων. / Taking into consideration that the use of computers has been transferred from research institutes to private companies and industries, we found out that computers constitute henceforth an integral enterprising tool. During the early 90’s, large private and public institutions afforded enormous quantities of data which revealed the need to be exploited. According to the above and having in mind that standing concepts handle information as the most important asset, it is noted that there is a need for applications that analyse and handle large amount of data. A solution to this problem was given by the data warehouse technologies and the analytical data processing (OLAP). Recently considerable research is conducted on ontologies and semantic networks, where information is already described in a conceptual way, which makes it easier to recover, to use it and compare to other similar. Thus, it is created the need of finding a new combination method of representation of knowledge. In the present master’s dissertation the reader is introduced into semantic networks. All the types of certain semantic networks are presented in detail and particularly those which are dealing with the representation of exceptions. Emphasis is given to the case of multidimensional databases as well as data deposits. Among them, OLAP tools and their basic theory is being described in more analysis. Finally, there is a presentation of a new technique that is based on methods for encountering exceptions in a semantic network whose goal is to extend one OLAP model in order to enable exception overcoming. The proposed expansion allows the existence of exceptions among the dimension values of a data hypercube which play a significant role to the right export of results.

No description available.

Presented work explores the possibilities of automatic construction and expansion of semantic networks with use of machine learning methods. The main focus is put on the feature retrieving procedure for the data set. The work presents a robust method of semantic relation retrieval, based on distributional hypothesis and trained on the data from Czech WordNet. We also show the first results for czech language in this area of research. Part of the thesis is also a set of software for processing and evaluating of input data and a overview and discussion about its results on real-world data. The resulting tools can process data of amount in orders of hundreds of millions of words. The research part of the thesis used Czech morphologicaly and syntacticaly annotated data, but the methods are not language dependent.