Natural Language Phrases Research Articles

Software expert discovery via knowledge domain embeddings in a collaborative network

Community Question Answering (CQA) websites can be claimed as the most major venues for knowledge sharing, and the most effective way of exchanging knowledge at present. Considering that massive amount of users are participating online and generating huge amount data, management of knowledge here systematically can be challenging. Expert recommendation is one of the major challenges, as it highlights users in CQA with potential expertise, which may help match unresolved questions with existing high quality answers while at the same time may help external services like human resource systems as another reference to evaluate their candidates. In this paper, we in this work we propose to exploring experts in CQA websites. We take advantage of recent distributed word representation technology to help summarize text chunks, and in a semantic view exploiting the relationships between natural language phrases to extract latent knowledge domains. By domains, the users’ expertise is determined on their historical performance, and a rank can be compute to given recommendation accordingly. In particular, Stack Overflow is chosen as our dataset to test and evaluate our work, where inclusive experiment shows our competence.

Searching for musical features using natural language queries: the C@merata evaluations at MediaEval

Musicological texts about classical music frequently include detailed technical discussions concerning the works being analysed. These references can be specific (e.g. C sharp in the treble clef) or general (fugal passage, Thor’s Hammer). Experts can usually identify the features in question in music scores but a means of performing this task automatically could be very useful for experts and beginners alike. Following work on textual question answering over many years as co-organisers of the QA tasks at the Cross Language Evaluation Forum, we decided in 2013 to propose a new type of task where the input would be a natural language phrase, together with a music score in MusicXML, and the required output would be one or more matching passages in the score. We report here on 3 years of the C@merata task at MediaEval. We describe the design of the task, the evaluation methods we devised for it, the approaches adopted by participant systems and the results obtained. Finally, we assess the progress which has been made in aligning natural language text with music and map out the main steps for the future. The novel aspects of this work are: (1) the task itself, linking musical references to actual music scores, (2) the evaluation methods we devised, based on modified versions of precision and recall, applied to demarcated musical passages, and (3) the progress which has been made in analysing and interpreting detailed technical references to music within texts.

Answering Natural Language Questions by Subgraph Matching over Knowledge Graphs

RDF question/answering (Q/A) allows users to ask questions in natural languages over a knowledge base represented by RDF. To answer a natural language question, the existing work takes a two-stage approach: question understanding and query evaluation. Their focus is on question understanding to deal with the disambiguation of the natural language phrases. The most common technique is the joint disambiguation, which has the exponential search space. In this paper, we propose a systematic framework to answer natural language questions over RDF repository (RDF Q/A) from a graph data-driven perspective. We propose a semantic query graph to model the query intention in the natural language question in a structural way, based on which, RDF Q/A is reduced to subgraph matching problem. More importantly, we resolve the ambiguity of natural language questions at the time when matches of query are found. The cost of disambiguation is saved if there are no matching found. More specifically, we propose two different frameworks to build the semantic query graph, one is relation (edge)-first and the other one is node-first. We compare our method with some state-of-the-art RDF Q/A systems in the benchmark dataset. Extensive experiments confirm that our method not only improves the precision but also speeds up query performance greatly.

The contribution of the lexical component in hybrid clustering, the case of four decades of “Scientometrics”

The introduction of textual analysis and the use of lexical similarities already proved an important asset in science mapping. Earlier research showed the added value of hybrid document networks over link-based ones through the reduction of the extreme sparseness. However, it was only after the application of Natural Language Processing and phrase extraction that networks purely based on lexical similarities could be used as input for topic detection in quantitative science studies. This study investigates the contribution of the lexical component in hybrid cluster on a set of articles published in the journal Scientometrics since its foundation during four decades. Shifting the weight of the lexical components generates changes in the structure of the underlying hybrid network, which can be detected through clustering techniques. We show that these changes are not moving documents randomly, but in fact identify small groups of papers either at the borderline between different topics or combining those. In addition, the analysis substantiates that the lexical component adopts the structure of the network rather than amplifies hidden structures of the link-based network.

Interpretation of the meaning of natural language phrases in problem-oriented systems

Interpretation of the meaning of natural language phrases in problem-oriented systems

Studies of Physician-Patient Communication with Older Patients: How Often is Hearing Loss Considered? A Systematic Literature Review.

Hearing loss is remarkably prevalent in the geriatric population: one-quarter of adults aged 60-69 and 80% of adults aged 80 years and older have bilateral disabling loss. Only about one in five adults with hearing loss wears a hearing aid, leaving many vulnerable to poor communication with healthcare providers. We quantified the extent to which hearing loss is mentioned in studies of physician-patient communication with older patients, and the degree to which hearing loss is incorporated into analyses and findings. We conducted a structured literature search within PubMed for original studies of physician-patient communication with older patients that were published since 2000, using the natural language phrase "older patient physician communication." We identified 409 papers in the initial search, and included 67 in this systematic review. Of the 67 papers, only 16 studies (23.9%) included any mention of hearing loss. In six of the 16 studies, hearing loss was mentioned only; in four studies, hearing loss was used as an exclusion criterion; and in two studies, the extent of hearing loss was measured and reported for the sample, with no further analysis. Three studies examined or reported on an association between hearing loss and the quality of physician-patient communication. One study included an intervention to temporarily mitigate hearing loss to improve communication. Less than one-quarter of studies of physician-elderly patient communication even mention that hearing loss may affect communication. Methodologically, this means that many studies may have omitted an important potential confounder. Perhaps more importantly, research in this field has largely overlooked a highly prevalent, important, and remediable influence on the quality of communication.

Expressing Biomedical Ontologies in Natural Language for Expert Evaluation.

We report on a study of our custom Hootation software for the purposes of assessing its ability to produce clear and accurate natural language phrases from axioms embedded in three biomedical ontologies. Using multiple domain experts and three discrete rating scales, we evaluated the tool on clarity of the natural language produced, fidelity of the natural language produced from the ontology to the axiom, and the fidelity of the domain knowledge represented by the axioms. Results show that Hootation provided relatively clear natural language equivalents for a select set of OWL axioms, although the clarity of statements hinges on the accuracy and representation of axioms in the ontology.

Socioeconomic gradients in the effects of universal school-based health behaviour interventions: a systematic review of intervention studies.

BackgroundSocioeconomic inequalities in health behaviour emerge in early life before tracking into adulthood. Many interventions to improve childhood health behaviours are delivered via schools, often targeting poorer areas. However, targeted approaches may fail to address inequalities within more affluent schools. Little is known about types of universal school-based interventions which make inequalities better or worse.MethodsSeven databases were searched using a range of natural language phrases, to identify trials and quasi-experimental evaluations of universal school-based interventions focused on smoking, alcohol, diet and/or physical activity, published from 2008–14. Articles which examined differential effects by socioeconomic status (N = 20) were synthesised using harvest plot methodology. Content analysis of 98 intervention studies examined potential reasons for attention or inattention to effects on inequality.ResultsSearches identified approximately 12,000 hits. Ninety-eight evaluations were identified, including 90 completed studies, of which 20 reported effects on SES inequality. There were substantial geographical biases in reporting of inequality, with only 1 of 23 completed North American studies testing differential effects, compared to 15 out of 52 completed European studies. Studies reported a range of positive, neutral or negative SES gradients in effects. All studies with a negative gradient in effect (i.e. which widened inequality) included educational components alone or in combination with environmental change or family involvement. All studies with positive gradients in effects included environmental change components, alone or combined with education. Effects of multi-level interventions on inequality were inconsistent. Content analyses indicated that in approximately 1 in 4 studies SES inequalities were discussed in defining the problem or rationale for intervention. Other potential barriers to testing effect on inequality included assumptions that universal delivery guaranteed universal effect, or that interventions would work better for poorer groups because they had most to gain.ConclusionsUniversal school-based interventions may narrow, widen or have no effect on inequality. There is a significant need for more routine testing of the effects of such interventions on inequality to enable firmer conclusions regarding types of interventions which affect inequality.PROSPERO registration numberCRD42014014548Electronic supplementary materialThe online version of this article (doi:10.1186/s12889-015-2244-x) contains supplementary material, which is available to authorized users.

Principles of developing an interactive system for the semantic processing of scientific and technical texts on chemical technology of reagents and ultrapure substances

The design and operating principles of an interactive system of semantic processing, i.e., understanding the meaning, of scientific and technical texts on chemical technology of reagents and ultrapure substances have been presented. Understanding of the meaning of scientific and technical texts has been implemented through the development of intelligent software of a dialogue system, which includes semantic and topological models of knowledge representation, as well as special languages of the automated generation of meaning. Understanding of the basic terms, concepts, and phrases in the limited natural language of the problem area has been presented in the form of the internal representation of the knowledge and language of control directives.

Հասկացումը որպես արհեստական ինտելեկտի ստեղծման առանցքային հիմնախնդիր

Հոդվածում քննարկվում են այն հանգա­մանքները, որոնց ազդեցությամբ համակար­գիչներով գերհագեցած այսօրվա իրողու­թյունում մոռացության է մատնվել ինքնու­րույն մտածող մեքենաների, արհեստական ինտելեկտի ստեղծման սկզբունքային խնդի­րը: Մտածող մեքենաների մշակման խնդիրը ձեռք բերեց գրեթե գործնական խնդրի կար­գավիճակ անցյալ դարի կեսից, երբ ստեղծ­վեցին առաջին էլեկտրոնային հաշվիչ մեքե­նաները, և իր բուռն ու կարճատև ծաղկումը ապրեց 20-րդ դարի նշանակալի նվաճում­ներից մեկը՝ կիբեռնետիկան: Սակայն բա­վականաչափ արագ պարզվեց, որ մեկ-երկու տասնամյակում մտածող մեքենաներ ստեղ­ծելու տարածված լավատեսական տրա­մադրվածությունը իրեն չի արդարացնում: Հոդվածի հեղինակը մատնանշում է արհես­տական ինտելեկտի ստեղծման երկու հիմ­նային դժվարություն՝ ինտելեկտուալ խնդիրների համընդհանուր ալգորիթմի մշակման խնդիրը և բնական լեզվով ներկա­յացված տեքստային տեղեկատվության հասկացման խնդիրը:
 Հոդվածում նշվում է, որ հեղինակին հաջողվել է մշակել մոտ 40 ենթակետից բաղկացած ինտելեկտուալ խնդիրների լուծ­ման համընդհանուր սխեմա, որը նա պայ­մանականորեն անվանվում է Խնդիրների լուծման համընդհանուր ալգորիթմ: Հեղի­նակը համոզված է, որ արհեստական ինտե­լեկտի ստեղծման համար անհրաժեշտ ին­տելեկտուալ խնդիրների լուծման համընդ­հանուր ալգորիթմը կարելի է ստանալ` հենց շարունակելով կոնկրետացնել իր կողմից մշակված սխեման: Հոդվածում ներկայաց­ված են Խնդիրների լուծման համընդհանուր ալգորիթմի հիմնական քայլերը:
 Համակարգիչների համար ստացվող տեղեկատվության ինքնուրույն հասկացման համար հեղինակն առաջարկում է հետևյալ միջոցները. բառերի տարբեր իմաստների համարակալում և համատեքստերի տիպերի ցուցիչների բացահայտում: Առաջարկվում է նաև խնդրի հասկացման մակարդակների դասա­կարգում:

Semantic standards and knowledge transfer in the problem of knowledge assessment on the basis of open tests

This paper is devoted to the problem of textual information transfer between experts and trainees in a computer-aided system for learning and testing of knowledge. A method is developed for the revelation of a semantic standard in a language usage situation (LUS) on the basis of word clustering by the occurrence frequency in semantically equivalent (SE) phrases in natural language (NL). An approach is presented to the revelation of stems and inflections in a word through the analysis of the occurrence frequency of letters in different positions with respect to the beginning and end of a word in the context of a LUS. An algorithm is developed for the extraction of structural units that correspond to syntagmas and determine the lexico-syntactic relations within a feature description of a LUS.

Google vs. the Library (Part II): Student Search Patterns and Behaviors when Using Google and a Federated Search Tool

This study examines the information-seeking behavior of undergraduate students within a research context. Student searches were recorded while the participants used Google and a library (federated) search tool to find sources (one book, two articles, and one other source of their choosing) for a selected topic. The undergraduates in this study believed themselves to be skilled researchers, but their search queries and behaviors did not support this belief. Students did not examine their topics to identify keywords and related terms. They relied heavily on the language presented to them via the list of research topics and performed natural language or simple keyword or phrase queries. They did not reformulate or refine their research questions or search queries, did not move beyond the first page of results, and did not examine metadata to improve searches. When using Google, students frequently visited commercial sites such as Amazon; content farms such as About.com ; and subscription databases such as JSTOR. This study concludes by offering suggestions for search interface improvement and pedagogical opportunities on which librarians may wish to focus or refocus. This article is the second in a series that examines student use of Google and a library (federated) search tool.

A new approach to query segmentation for relevance ranking in web search

In this paper, we try to determine how best to improve state-of-the-art methods for relevance ranking in web searching by query segmentation. Query segmentation is meant to separate the input query into segments, typically natural language phrases. We propose employing the re-ranking approach in query segmentation, which first employs a generative model to create the top k candidates and then employs a discriminative model to re-rank the candidates to obtain the final segmentation result. The method has been widely utilized for structure prediction in natural language processing, but has not been applied to query segmentation, as far as we know. Furthermore, we propose a new method for using the results of query segmentation in relevance ranking, which takes both the original query words and the segmented query phrases as units of query representation. We investigate whether our method can improve three relevance models, namely n-gram BM25, key n-gram model and term dependency model, within the framework of learning to rank. Our experimental results on large scale web search datasets show that our method can indeed significantly improve relevance ranking in all three cases.

Spatial Extent Models for Natural Language Phrases Involving Directional Containment

AbstractWe study the problem of assigning a spatial extent to a text phrase such as ‘central northern California’, with the objective of allowing spatial interpretations of natural language, and consistency testing of complex utterances that involve multiple phrases from which spatial extent can be derived. The work presented addresses alternative spatial extent assignations, and the evaluation of those alternatives to decide on the best performing ones. Three important classes of these phrases are identified and we call them full direction phrases, half‐direction phrases, and extreme direction phrases, respectively. Examples of these are: ‘north‐western California’, ‘central northern California’, and ‘extreme south‐eastern California’. Such descriptions vaguely delineate a spatial extent, and we attempt to derive a region‐representation for them. Our approach is to identify a number of extent assignation parameters, which are used to derive a variety of spatial extent models, each of which we evaluate against a large gazetteer corpus. Three example assignation parameters used are: choice of center, choice of shape of central sector, and choice of shape of outer sectors. The corpus used for the evaluation is the two‐volume Ornithological Gazetteer of Brazil (Paynter and Traylor 1991a, b). It allows us to compare the spatial extent assignations and derive recommendations for such assignations for future use.

An evaluation of retranslation methods in computing with words

To represent output fuzzy values of a computing with words (CW) system in natural language, a retranslation unit is required. In this work, retranslation methods applicable to a CW system are explored. Several methods that employ similarity measures of fuzzy sets, linguistic modifiers, or linguistic quantifiers have been applied to three real-world case studies. Performances of the applied methods have been evaluated through degree of validity, and comparison of characteristics of fuzzy sets such as fuzziness and specificity. Results show that invalid linguistic terms might be used in the retranslation process which also cause incomprehensible phrases in natural language.

Evolving Grounded Spatial Language Strategies

Each natural language phrase is evidence for a particular strategy of construing reality. One domain where this has been extensively studied is spatial language, which reveals an enormous amount of variation of conceptualization strategies both within a particular language and cross-culturally. This paper proposes a computational formalism for representing conceptualization strategies and shows how the formalism can be used to study and explain the evolution and emergence of spatial conceptualization strategies and their impact on shared grounded communication systems.

Requirements-level language and tools for capturing software system essence

Creation of an unambiguous requirements specification with precise domain vocabulary is crucial for capturing the essence of any software system, either when developing a new system or when recovering knowledge from a legacy one. Software specifications usually maintain noun notions and include them in central vocabularies. Verb or adjective phrases are easily forgotten and their definitions buried inside imprecise paragraphs of text. This paper proposes a model-based language for comprehensive treatment of domain knowledge, expressed through constrained natural language phrases that are grouped by nouns and include verbs, adjectives and prepositions. In this language, vocabularies can be formulated to describe behavioural characteristics of a given problem domain. What is important, these characteristics can be linked from within other specifications similarly to a wiki. The application logic can be formulated through sequences of imperative subject-predicate sentences containing only links to the phrases in the vocabulary. The paper presents an advanced tooling framework to capture application logic specifications making them available for automated transformations down to code. The tools were validated through a controlled experiment.

An inference engine toolkit for computing with words

Computing with Words is an emerging paradigm in knowledge representation and information processing. It provides a mathematical model to represent the meaning of imprecise words and phrases in natural language and introduces advanced techniques to perform reasoning on inexact knowledge. Since its introduction, there have been many studies on computing with words but mostly from the theoretical point of view and the paradigm still lacks sufficient support from the software side. This paper is an attempt to fill this gap by presenting an enhanced inference engine toolkit for supporting computing with words. The scope of the presented toolkit, as opposed to many available fuzzy logic tools, goes beyond simple fuzzy-if-then rules and performs a chain of inferences on complex fuzzy propositions containing fuzzy arithmetics, fuzzy quantifiers, and fuzzy probabilities. The toolkit may be appealing to researchers, practitioners, and educators in knowledge based applications and soft computing as it implements a powerful declarative language which allows users to express their knowledge in a more natural and convenient way and performs a chain of reasoning on imprecise propositions.

Protocol for a systematic review of the effects of schools and school-environment interventions on health: evidence mapping and syntheses.

BackgroundSchools may have important effects on students' and staff's health. Rather than treating schools merely as sites for health education, 'school-environment' interventions treat schools as settings which influence health. Evidence concerning the effects of such interventions has not been recently synthesised.Methods/designSystematic review aiming to map and synthesise evidence on what theories and conceptual frameworks are most commonly used to inform school-environment interventions or explain school-level influences on health; what effects school-environment interventions have on health/health inequalities; how feasible and acceptable are school-environment interventions; what effects other school-level factors have on health; and through what processes school-level influences affect health.We will examine interventions aiming to promote health by modifying schools' physical, social or cultural environment via actions focused on school policies and practices relating to education, pastoral care and other aspects of schools beyond merely providing health education. Participants are staff and students age 4-18 years.We will review published research unrestricted by language, year or source. Searching will involve electronic databases including Embase, ERIC, PubMed, PsycInfo and Social Science Citation Index using natural-language phrases plus reference/citation checking.Stage 1 will map studies descriptively by focus and methods. Stage 2 will involve additional inclusion criteria, quality assessment and data extraction undertaken by two reviewers in parallel. Evidence will be synthesised narratively and statistically where appropriate (undertaking subgroup analyses and meta-regression and where no significant heterogeneity of effect sizes is found, pooling these to calculate a final effect size).DiscussionWe anticipate: finding a large number of studies missed by previous reviews; that non-intervention studies of school effects examine a greater breadth of determinants than are addressed by intervention studies; and that intervention effect estimates are greater than for school-based health curriculum interventions without school-environment components.

A modal learning adaptive function neural network applied to handwritten digit recognition

A novel combination of the adaptive function neural network (ADFUNN) and on-line snap–drift learning is presented in this paper and applied to optical and pen-based recognition of handwritten digits [E. Alpaydin, F. Alimoglu for Optical Recognition of Handwritten Digits and E. Alpaydin, C. Kaynak for Pen-Based Recognition of Handwritten Digits http://www.ics.uci.edu/~mlearn/databases/optdigits/ http://www.ics.uci.edu/~mlearn/databases/pendigits/]. Snap–drift [S.W. Lee, D. Palmer-Brown, C.M. Roadknight, Performance-guided neural network for rapidly self-organising active network management (Invited Paper), Journal of Neurocomputing, 61C, 2004, pp. 5–20] employs the complementary concepts of common (intersection) feature learning (called snap) and LVQ (drift towards the input patterns) learning, and is a fast, unsupervised method suitable for on-line learning and non-stationary environments where new patterns are continually introduced. ADFUNN [M. Kang, D. Palmer-Brown, An adaptive function neural network (ADFUNN) for phrase recognition, in: The International Joint Conference on Neural Networks (IJCNN05), Montréal, Canada, 2005, D. Palmer-Brown, M. Kang, ADFUNN: An adaptive function neural network, in: The 7th International Conference on Adaptive and Natural Computing Algorithms (ICANNGA05), Coimbra, Portugal, 2005] is based on a linear piecewise neuron activation function that is modified by a novel gradient descent supervised learning algorithm. It has recently been applied to the Iris dataset, and a natural language phrase recognition problem, exhibiting impressive generalisation classification ability with no hidden neurons. The unsupervised single layer snap–drift is effective in extracting distinct features from the complex cursive-letter datasets, and the supervised single layer ADFUNN is capable of solving linearly inseparable problems rapidly. In combination within one network (SADFUNN), these two methods are more powerful and yet simpler than MLPs, at least on this problem domain. We experiment on SADFUNN with two handwritten digits datasets problems from the UCI Machine Learning repository. The problems are learned rapidly and higher generalisation results are achieved than with a MLP.