Теории воплощенного познания в задаче формальной репрезентации семантики на основе мультиагентных нейрокогнитивных архитектур

Developing software systems is a process of eliciting and evolving stakeholder requirements using informal, semiformal and formal representations of knowledge. Managing the transitions between these representations is a challenging task. On the one hand informal representations for capturing system requirements support straightforward involvement and interaction of even inexperienced stakeholders. On the other hand this approach may result in inconsistent, incomplete, and ambiguous specifications. Semiformal and formal representations overcome these problems but are less suited for stakeholder interaction. In this paper, we describe an experiment using the WinWin model for negotiating system requirements and the Unified Modelling Language. We discuss our experiences and draw conclusions useful in bridging informal and formal representations.

Este artigo visa analisar os princípios de seleção do modo verbal no ámbito do sub-sistema doxástico no português do Brasil. Especialmente, esse estudo, que combina uma reflexão teórica sobre a modalidade doxástica com uma pesquisa empírica com base no Corpus do Português de Davies & Ferreira, pretende discutir a semântica de contextos intensionais criados por verbos de crença desde uma perspetiva formal, proposta da semântica modal. Para poder situar o sub-sistema do português do Brasil dentro do panorama pan-románico, estabeleceremos as diferenças entre as diferentes línguas románicas no que tange os princípios de seleção do modo verbal em contextos doxásticos. Apresentaremos os princípios particulares de seleção do modo verbal em cada língua románica numa formalização adequada.

The research aims to develop a procedure for the formal representation of literary text semantics through a domain ontology, using the novella “Monday Begins on Saturday” by Arkady and Boris Strugatsky as a case study. In this article, ontologization is treated as the translation of semantic units from the fictional world (actants, objects, spaces, and events) and the relationships between them into a formal RDF/OWL model linked to specific textual nominations. The object of the study is the semantic organization of a literary text, while the subject is the ontological modeling of narrative entities and events along with their typed connections. The scientific novelty lies in the development and description of a step-by-step (operational) protocol for the ontologization of literary texts – ranging from the fragmentation and recording of entities/events and their links in tabular format to the construction of a full RDF/OWL model. The results demonstrate that the proposed procedure (fragmenting the text into episodes, tabular recording of instances with unique identifiers, and building the RDF/OWL ontology in Protege) allows for translating research questions into verifiable queries and obtaining reproducible samples using SPARQL. For example, it enables the extraction of linked episodes, the reconstruction of movement trajectories, and the comparison of role configurations and event types. A domain ontology of the fictional world was created, identifying and classifying 836 units. This demonstrates the feasibility of the proposed approach for further scaling to mini-corpora of plot-related works and expanded models of authorial universes.

Default Semantics (DS) is a theory of discourse that represents the main meaning intended by the speaker and recovered by the addressee, using truth-conditional, formal, but pragmatics-rich representations. It was originally developed at the University of Cambridge by K. M. Jaszczolt in the late 1990s and has since been applied to a variety of constructions, phenomena, and languages. The category of primary meaning, as it is understood in DS and represented in its semantic qua conceptual representations, cuts across the explicit/implicit divide. Semantic representations are not limited by the constraints imposed by the logical form of the sentence; they allow for its modifications but also, unlike in other post-Gricean theories, they allow for it to be overridden when the main informational content is conveyed through sources other than the linguistic expression itself. DS identifies five sources of information, all operating on an equal footing: word meaning and sentence structure (WS); situation of discourse (SD); properties of human inferential system (IS); stereotypes and presumptions about society and culture (SC); and world knowledge (WK). Since all of the sources can contribute to the truth-conditional representation, the traditional syntactic constraint that ties the representation to the logical form of the uttered sentence could be abandoned, resulting in modeling a cognitively plausible, main message as intended by a Model Speaker and recovered by a Model Addressee. As a result, DS-theoretic representations can pertain either to (i) the logical form of the sentence; (ii) the logical form with saturated indexical expressions; (iii) the logical form that is freely modified; as well as (iv) representations that do not make use of the logical form of the sentence. The identified sources of information are mapped onto four types of processes that interact in producing the representation (called merger representation, or Σ): processing of word meaning and sentence structure (WS); conscious pragmatic inference (CPI); cognitive defaults (CD, capturing strong informativeness, or strong intentionality of the underlying mental states, for example referential rather attributive reading of definite descriptions); and social, cultural, and world knowledge defaults (SCWD, capturing the relevant sociocultural conventions and encyclopedic knowledge). “Defaults” are understood there as automatic interpretations, “shortcuts through conscious inference” for the speaker and for the context, and as such are by definition not cancellable and are immune to controversies engendered by the “noncism”-“defaultism” debates in post-Gricean pragmatics. DS subscribes to the methodological and ontological assumption of compositionality of meaning on the level of such cognitive representations (Σs).

Acquisition and validation of software requirements

One of the strong points of E-commerce websites is that they are often abundant with product reviews from consumers who experienced the products and testify to the usefulness of the products or otherwise. These reviews are helpful for consumers to optimize their purchasing decisions. However, while popular products receive many reviews, many other products do not have an adequate number of reviews leading to the cold item problem. In this proposal, we propose a solution outline for the cold item problem by automatically generating reviews and predicting ratings for the cold products from available reviews of similar products in e-commerce websites as well as users' opinion shared in the microblogging platforms such as Twitter. We propose a framework to build a formal semantic representation of products from unstructured product descriptions, user reviews as well as user ratings. Such presentations assist us to measure product similarity and relatedness in a accurate and cost-effective way. Besides, we propose a model to generate additional reviews for a cold product by mining users' posts shared on medium such as Twitter and transfer them to the e-commerce website. Preliminary experiments show promising results in finding products similar to the cold products.

Formal verification of business process models can be done through model checking (also known as property checking), where a model checker tool may automatically find violations of properties in a process model. This approach obviously has formal representations as a prerequisite. However, a key challenge for applying this approach in practice is to consistently formalize the process and its properties, which clearly cannot be done automatically. We studied this challenge in a case study of formally verifying an informally given business process against a guideline written like a legal text. Major lessons learned from this case study are that formalizing is key to success and that in its course a semi-formal representation of properties is useful. In the course of such a step-wise and incremental formalization, problems with the given process model have been found already, apart from those found with a model checker tool that used the formal property specification. In total, our approach revealed five problems not found by the official review. In summary, this paper investigates in a case study consistently formalizing a business process and its properties for verification through model checking.

With the rapid evolution of web technology and Internet, the mathematical web has emerged as a new research trend, which changes the traditional manner of mathematical scientific researching, mathematical production sharing, mathematical knowledge engineering and mathematical education. In addition, the acquisition of mathematical knowledge and the integration of diversiform of mathematical software are also taken into considers on the web environment. In this paper, we will present a web-based, user adaptive learning and grid computing environment that employs content encoded in a semantic representation of mathematical knowledge and integrates mathematics software as services. We have brought forward a concept which makes the mathematical computing environment working more efficiently. This mathematics computing environment mainly offers a powerful and convenient mathematics computing environment for end users and service providers. From the view of computing, this environment is a web-based service-oriented environment, built on three-layered resource model. From the view of end users, the formal representation allows semantics-based retrieval of distributed mathematical facts, giving mathematicians a tool and an easy-to-learn mathematical language to access and query the present mathematical knowledge and services. On the other hand, we use grid toolkits, which have the capabilities to aggregate and coordinate the computational power of geographically distributed, heterogeneous and dynamic resources efficiently, to organize and manage the mathematical resources.

To make legal texts machine processable, the texts may be represented as linked documents, semantically tagged text, or translated to formal representations that can be automatically reasoned with. The paper considers the latter, which is key to testing consistency of laws, drawing inferences, and providing explanations relative to input. To translate laws to a form that can be reasoned with by a computer, sentences must be parsed and formally represented. The paper presents the state-of-the-art in automatic translation of law to a machine readable formal representation, provides corpora, outlines some key problems, and proposes tasks to address the problems.

While sequence-to-sequence (seq2seq) models achieve state-of-the-art performance in many natural language processing tasks, they can be too slow for real-time applications. One performance bottleneck is predicting the most likely next token over a large vocabulary; methods to circumvent this bottleneck are a current research topic. We focus specifically on using seq2seq models for semantic parsing, where we observe that grammars often exist which specify valid formal representations of utterance semantics. By developing a generic approach for restricting the predictions of a seq2seq model to grammatically permissible continuations, we arrive at a widely applicable technique for speeding up semantic parsing. The technique leads to a 74% speed-up on an in-house dataset with a large vocabulary, compared to the same neural model without grammatical restrictions.

To obtain formal semantics of Chinese discourse, compositional discourse representation theory was adopted to compose semantics from Chinese word to sentence and to discourse. According to syntactic rules and λ-box expressions of word categories, sentential logic expression could be created by translating from child nodes to father node in syntactic tree, then the formal representation of Chinese discourse could be further obtained by using construction procedure of DRT to resolve cross-sentential demonstrates and merge sentential DRSs. A rough prototype system was given to depict the implementment process, which will benefit establishing machine-inside representation of discourse semantics of contemporary Chinese.Keywordscompositional discourse representation theoryChinese discourseformal semantics

Meaning Representation formalism is rapidly emerging as an important practical form of structured sentence semantics which, thanks to the availability of largescale annotated corpora, has potential as a convergence point for NLP research.This tutorial unmasks the design philosophy, data creation process, and existing algorithms for AMR semantics.It is intended for anyone interested in working with AMR data, including parsing text into AMRs, generating text from AMRs, and applying AMRs to tasks such as machine translation and summarization.

The fourth industrial revolution demands flexibility, adaptability, transparency and semantic interoperability. Within the German Industry 4.0 initiative, the Reference Architecture Model Industrie 4.0 (RAMI4.0) has recently been standardized and OPC Unified Architecture (OPC UA) is listed as the sole recommendation for implementation of a communication layer. Even though OPC UA helps bridge the interoperability gap at the automation level, its semantic has not yet been formally defined and an efficient automated reasoning and knowledge inference on the OPC UA address space is therefore not yet possible. This paper addresses this issue by presenting a solution to infer knowledge from OPC UA information models. By analyzing and comparing the semantic expressiveness of the OPC UA address space with semantic knowledge representation formalisms such as RDF and OWL, we derived and implemented a solution to transform an OPC UA information model into an RDF-Graph expressing an OWL Ontology. Using the generated ontology, we were able to run a true reasoning task directly on the OPC UA address space. In this paper, the developed approach is conveniently validated on various case studies involving online factory reconfiguration, intelligent energy management, and human-machine interfaces.

A unified model of acting and inference (called rational engine) for an intelligent architecture is presented. The integrated architecture uses propositional semantic network representations of beliefs, goals, acts, plans, and deduction rules. A message passing scheme implements the underlying logic, and has forward, backward, and bidirectional inference capabilities. It also incorporates the notions of truth maintenance, spreading activation, and a unified representational formalism for acting and inference called transformers. >

This paper describes an application for computing first-order semantic representations of English texts. It is based on a combination of hybrid shallow-deep components arranged within the middleware framework Heart of Gold. The shallow-deep semantic analysis employs Robust Minimal Recursion Semantics (RMRS) as a common semantic underspecification formalism for natural language processing components. In order to compute efficiently first-order representations of the input text, the intermediate RMRS results of the shallow-deep analysis are transformed into the dominance constraints formalism and resolved by the underspecification resolver UTool. First-order expressions can serve as a formal knowledge representation of natural text and thus can be utilized in knowledge engineering or textual reasoning. At the end of this paper, we describe their application for recognizing textual entailment.