Natural Language Analysis Research Articles

A Quantitative Meta-analysis of Neuroimaging Studies of Pragmatic Language Comprehension: In Search of a Universal Neural Substrate

Pragmatics may be defined as the ability to communicate by expressing and recognizing intentions. The objective of this meta-analysis was to identify neural substrates for comprehension of pragmatic content in general, as well as the differences between pragmatic forms, and to describe if there is differential recruitment of brain areas according to natural language. This meta-analysis included 48 functional magnetic resonance imaging studies that reported pragmatic versus literal language contrasts. The pragmatic forms were speech acts, metaphors, idioms, and irony. Effect Size-Signed Differential Mapping software was used to calculate the mean for all contrasts as well as for each pragmatic form, and make comparisons among all forms. Due to variations in pragmatic content configuration such as natural language, stimulus modality, and writing systems, these variations were also analyzed with subgroups’ analyses. The analyses found a highly reproducible bilateral fronto-temporal and medial prefrontal cortex network for pragmatic comprehension. Each pragmatic form showed a specific convergence pattern within this bilateral network. Natural language analyses showed that fronto-temporal regions were recruited by Germanic languages, while only left frontal areas were recruited by Romance languages, and right medial prefrontal cortex by Japanese. In conclusion, pragmatic language comprehension involves classical language areas in bilateral perisylvian regions, along with the medial prefrontal cortex, an area involved in social cognition. Together, these areas could represent the “pragmatic language network”. Nonetheless, when proposing a universal neural substrate for all forms of pragmatic language, the diversity among studies in terms of pragmatic form, and configuration, must be taken into consideration.

Parsing Arabic Verb Phrases Using Pregroup Grammars

Parsing of Arabic phrases is a crucial requirement for many applications such as question answering and machine translation. The calculus of pregroup introduced by Lambek as an algebraic computational machinery for the grammatical analysis of natural languages. Pregroup grammar used to analyse sentence structure in many European languages such as English, and non-European languages such as Japanese. In Arabic language, Lambek employed the notions of pregroup to analyse some grammatical structures such as conjugating the verb, tense modifiers and equational sentences. This work attempts to develop an initial phase of an efficient automatic pregroup grammar parser by using linear approach to analysethe verbal phrases of Modern Standard Arabic (MSA). The proposed system starts building Arabic lexicon contains all possible categories of Arabic verbs, then analysing the input verbal Arabic phrase to check if it is wellformed by using linear parsing algorithm based on pregroup grammar rules.

A Social Media Based Examination of the Effects of Counseling Recommendations after Student Deaths on College Campuses

Student deaths on college campuses, whether brought about by a suicide or an uncontrollable incident, have serious repercussions for the mental wellbeing of students. Consequently, many campus administrators implement post-crisis intervention measures to promote student-centric mental health support. Information about these measures, which we refer to as "counseling recommendations", are often shared via electronic channels, including social media. However, the current ability to assess the effects of these recommendations on post-crisis psychological states is limited. We propose a causal analysis framework to examine the effects of these counseling recommendations after student deaths. We leverage a dataset from 174 Reddit campus communities and ~400M posts of ~350K users. Then we employ statistical modeling and natural language analysis to quantify the psychosocial shifts in behavioral, cognitive, and affective expression of grief in individuals who are "exposed" to (comment on) the counseling recommendations, compared to that in a matched control cohort. Drawing on crisis and psychology research, we find that the exposed individuals show greater grief, psycholinguistic, and social expressiveness, providing evidence of a healing response to crisis and thereby positive psychological effects of the counseling recommendations. We discuss the implications of our work in supporting post-crisis rehabilitation and intervention efforts on college campuses.

Logical analysis of empirical expressions. What is wrong with empiricism

Abstract The following well-known problem motivated my handling more general problems. As we surely know, our pupils and even students are confronted with much more trouble when learning mathematics (and even physics) than when they learn ‘empirical’ sciences like biology, mineralogy etc. There are many factors that can at least partially explain this phenomenon. I would however mention one factor that is not too frequently adduced: mathematics, logic, and much of physics use concepts that are abstract while the empirical sciences seem to support understanding by using expressions concerning (denoting? expressing?) concrete objects. Therefore the first topic to be explained (or explicated) is: Abstract vs. concrete. The second point will consist of applying the first point to explanation of the trouble with learning mathematics. The third point will ask Logical Analysis of Natural Language how to tell abstract expressions from concrete ones. The fourth point will confront the concept described in the foregoing point with conceptions trying to abandon the distinction between analytic and empirical expressions. Here it will be shown that the empiricism representing this latter conception deprives semantics as applied to Natural language of important features of expressivity.

Exploring the Potentials of ICT Tools for Human-Centric Regenerative Design

Aiming for sustainable buildings and cities is critical to achieving a future that is socially just, ecologically regenerative, culturally rich, and economically viable. However, our current concepts of sustainability often exclude the essential domains of data, information, and the knowledge relating to the relationship between buildings and people that inhabit them. Thus, the research questions at the core of this paper have been as follows: Can technologies and artificial intelligence (AI) be used to create systems that enhance relationships between buildings and inhabitants? Can social networks and natural interactions support further research relating to human-centric design tools for the built environment? The Human Observation Meta-Environment (HOME) project was developed to address this question. The ICT architecture has been tested to observe and collect human behaviour data within a sentient room at the Politecnico di Torino (IT), where the inhabitants were strategically aware of their behaviours. Methods of analysis included technologies related to the domain of AI (such as Natural Language Analysis, Computer Vision, Machine Learning and Deep Learning) that have been used in social network analysis in connection with the word ‘comfort’, and definitions resonate strongly with the realm of regenerative design. Results were used to further research the role of users that could serve as leverages to design (both spaces and related smart systems) according to actual user needs. People from very different disciplinary backgrounds interacted with the prototype in a workshop and provided stimuli for further considerations regarding the possible technological, psychological, cognitive, cultural, social, political, and aesthetical impacts of the use of these technologies inside sentient buildings. The paper enriches the discourse on how ICT data can be organised and read in a human-centric regenerative design process perspective.

Decoding the Musical Message via the Structural Analogy between Verbal and Musical Language

Abstract The topic approached in this paper aims to identify the structural similarities between the verbal and the musical language and to highlight the process of decoding the musical message through the structural analogy between them. The process of musical perception and musical decoding involves physiological, psychological and aesthetic phenomena. Besides receiving the sound waves, it implies complex cognitive processes being activated, whose aim is to decode the musical material at cerebral level. Starting from the research methods in cognitive psychology, music researchers redefine the process of musical perception in a series of papers in musical cognitive psychology. In the case of the analogy between language and music, deciphering the musical structure and its perception are due, according to researchers, to several common structural configurations. A significant model for the description of the musical structure is Noam Chomsky’s generative-transformational model. This claimed that, at a deep level, all languages have the same syntactic structure, on account of innate anatomical and physiological structures which became specialized as a consequence of the universal nature of certain mechanisms of the human intellect. Chomsky’s studies supported by sophisticated experimental devices, computerised analyses and algorithmic models have identified the syntax of the musical message, as well as the rules and principles that underlie the processing of sound-related information by the listener; this syntax, principles and rules show surprising similarities with the verbal language. The musicologist Heinrich Schenker, 20 years ahead of Chomsky, considers that there is a parallel between the analysis of natural language and that of the musical structure, and has developed his own theory on the structure of music. Schenker’s structural analysis is based on the idea that tonal music is organized hierarchically, in a layering of structural levels. Thus, spoken language and music are governed by common rules: phonology, syntax and semantics. Fred Lerdahl and Ray Jackendoff develop a musical grammar where a set of generating rules are defined to explain the hierarchical structure of tonal music. The authors of the generative theory propose the hypothesis of a musical grammar based on two types of rules, which take into account the conscious and unconscious principles that govern the organization of the musical perception. The structural analogy between verbal and musical language consists of several common elements. Among those is the hierarchical organization of both fields, a governance by the same rules – phonology, syntax, semantics – and as a consequence of the universal nature of certain mechanisms of the human intellect, decoding the transmitted message is accomplished thanks to some universal innate structures, biologically inherited. Also, according to Chomsky's linguistics model a musical grammar is configured, one governed by wellformed rules and preference rules. Thus, a musical piece is not perceived as a stream of disordered sounds, but it is deconstructed, developed and assimilated at cerebral level by means of cognitive pre-existing schemes.

Natural speech algorithm applied to baseline interview data can predict which patients will respond to psilocybin for treatment-resistant depression

BackgroundNatural speech analytics has seen some improvements over recent years, and this has opened a window for objective and quantitative diagnosis in psychiatry. Here, we used a machine learning algorithm applied to natural speech to ask whether language properties measured before psilocybin for treatment-resistant can predict for which patients it will be effective and for which it will not. MethodsA baseline autobiographical memory interview was conducted and transcribed. Patients with treatment-resistant depression received 2 doses of psilocybin, 10 mg and 25 mg, 7 days apart. Psychological support was provided before, during and after all dosing sessions. Quantitative speech measures were applied to the interview data from 17 patients and 18 untreated age-matched healthy control subjects. A machine learning algorithm was used to classify between controls and patients and predict treatment response. ResultsSpeech analytics and machine learning successfully differentiated depressed patients from healthy controls and identified treatment responders from non-responders with a significant level of 85% of accuracy (75% precision). ConclusionsAutomatic natural language analysis was used to predict effective response to treatment with psilocybin, suggesting that these tools offer a highly cost-effective facility for screening individuals for treatment suitability and sensitivity. LimitationsThe sample size was small and replication is required to strengthen inferences on these results.

“I’m 13. I’m online. U believe me?”: Implications for undercover Internet stings.

We examined the extent to which 262 adults, recruited from both a U.S. university sample and Amazon’s Mechanical Turk, believed the identity of a 13-year-old boy or girl persona in an online chat room. Sixteen undergraduate confederates (aged 19-38) followed two basic biographical sketches (Amber or James) and were instructed to “chat like a teenager” in one-to-one private chats. Confederates cycled through conditions where they provided no age or gender information (control), stated their age and gender, stated their age and gender and provided a picture (attractive or average), or stated their grade only. In all but the control condition, participants estimated the average age of their chat partner to be 13-14. Meanwhile, when confederates stated their age and gender, 83-88% of participants believed the confederate’s stated age, and 94-98% believed their gender. When asked about cues used to discern age and gender, most participants (98%) used multiple cues, including stated age and gender, content cues, style cues, and picture. Moreover, natural language analyses showed that the confederates used significantly fewer analytic words, fewer six-letter words, and displayed less clout than the adults with whom they chatted, and confederates who used more analytic words were perceived as older. Our findings contradict Lincoln and Coyle (2013) and suggest that even confederates who are not specially trained law enforcement agents are able to deceive others as to their age and gender in online chat rooms. These findings have important implications for those conducting undercover Internet stings and those who prosecute those cases.

Reconstructing news spread networks and studying its dynamics

News spread in internet media outlets can be seen as a contagious process forming temporal networks representing the influence between published articles. In this article we propose a methodology based on the application of natural language analysis of the articles to reconstruct the news spread network. From the reconstructed network, we show that the dynamics of the news spread can be approximated by a classical SIR epidemiological dynamics upon the network.

Semi-automatic derivation of RESTful choreographies from business process choreographies

Enterprises reach out for collaborations with other organizations in order to offer complex products and services to the market. Such collaboration and coordination between different organizations, for a good share, is facilitated by information technology. The BPMN process choreography is a modeling language for specifying the exchange of information and services between different organizations at the business level. Recently, there is a surging use of the REST architectural style for the provisioning of services on the web, but few systematic engineering approach to design their collaboration. In this paper, we address this gap in a comprehensive way by defining a semi-automatic method for the derivation of RESTful choreographies from process choreographies. The method is based on natural language analysis techniques to derive interactions from the textual information in process choreographies. The proposed method is evaluated in terms of effectiveness resulting in the intervention of a web engineer in only about 10% of all generated RESTful interactions.

Parsing Arabic Verb Phrases Using Pregroup Grammars

Parsing of Arabic phrases is a crucial requirement for many applications such as question answering and machine translation. The calculus of pregroup introduced by Lambek as an algebraic computational machinery for the grammatical analysis of natural languages. Pregroup grammar used to analyse sentence structure in many European languages such as English, and non-European languages such as Japanese. In Arabic language, Lambek employed the notions of pregroup to analyse some grammatical structures such as conjugating the verb, tense modifiers and equational sentences. This work attempts to develop an initial phase of an efficient automatic pregroup grammar parser by using linear approach to analysethe verbal phrases of Modern Standard Arabic (MSA). The proposed system starts building Arabic lexicon contains all possible categories of Arabic verbs, then analysing the input verbal Arabic phrase to check if it is wellformed by using linear parsing algorithm based on pregroup grammar rules.

Perfect ratings with negative comments: Learning from contradictory patient survey responses

This research explores why patients give perfect domain scores yet provide negative comments on surveys. In order to explore this phenomenon, vendor-supplied in-patient survey data from eleven different hospitals of a major U.S. health care system were utilized. The dataset included survey scores and comments from 56,900 patients, collected from January 2015 through October 2016. Of the total number of responses, 30,485 (54%) contained at least one comment. For our analysis, we use a two-step approach: a quantitative analysis on the domain scores augmented by a qualitative text analysis of patients’ comments. To focus the research, we start by building a hospital recommendation model using logistic regression that predicts a patient’s likelihood to recommend the hospital; we use this to further evaluate the top four most predictive domains. In these domains (personal issues, nurses, hospital room, and physicians), a significant percentage of patients who rated their experience with a perfect domain score left a comment categorized as not positive, thus giving rise to stark contrasts between survey scores and comments provided by patients. Within each domain, natural language analysis of patient comments shows that, despite providing perfect survey scores, patients have much to say to health care organizations about their experiences in the hospital. A summary of comments also shows that respondents provide negative comments on issues that are outside the survey domains. Results confirm that harvesting and analyzing comments from these patients is important, because much can be learned from their narratives. Implications for health care professionals and organizations are discussed.

Crowdsourcing a Parallel Corpus for Conceptual Analysis of Natural Language

Computer users today are demanding greater performance from systems that understand and respond intelligently to human language as input. In the past, researchers proposed and built conceptual analysis systems that attempted to understand language in depth by decomposing a text into structures representing complex combinations of primitive acts, events, and state changes in the world the way people conceive them. However, these systems have traditionally been time-consuming and costly to build and maintain by hand. This paper presents two studies of crowdsourcing a parallel corpus to build conceptual analysis systems through machine learning. In the first study, we found that crowdworkers can view simple English sentences built around specific action words, and build conceptual structures that represent decompositions of the meaning of that action word into simple and complex combinations of conceptual primitives. The conceptual structures created by crowdworkers largely agree with a set of gold standard conceptual structures built by experts, but are often missing parts of the gold standard conceptualization. In the second study, we developed and tested a novel method for improving the corpus through a subsequent round of crowdsourcing; In this “refinement” step, we presented only conceptual structures to a second set of crowdworkers, and found that when crowdworkers could identify the action word in the original sentence based only on the conceptual structure, the conceptual structure was a stronger match to the gold standard structure for that sentence. We also calculated a statistically significant correlation between the number of crowdworkers who identified the original action word for a conceptual structure, and the degree of matching between the conceptual structure and a gold standard conceptual structure. This indicates that crowdsourcing may be used not only to generate the conceptual structures, but also to select only those of the highest quality for a parallel corpus linking them to natural language.

Recruitment in a Crowdsource Environment: An Exploratory Investigation

We conduct an exploratory study to investigate recruitment practice in the context of nonstandard project-based jobs in a crowdsource environment, where organizations utilize contributions from online internet communities of users to fulfill needs services or ideas. Crowdsourcing platforms have become omnipresent in today’s globally connected marketplace, and have become important sources for specialized talent in a variety of idea-driven job tasks and projects. Employing novel techniques associated with text-mining and natural language analysis, our findings suggest that factors associated with both inducements, and the amount and structure of information contained in the job advertisement posting influence a range of aggregate variables that include the size and quality of the applicant pool. Results are discussed both in terms of their implications for how talent is sourced in such virtual environments, as well as how current knowledge of recruitment can be expanded to accommodate nonstandard jobs.

Artificial Intelligence-Assisted Online Social Therapy for Youth Mental Health.

Introduction: Benefits from mental health early interventions may not be sustained over time, and longer-term intervention programs may be required to maintain early clinical gains. However, due to the high intensity of face-to-face early intervention treatments, this may not be feasible. Adjunctive internet-based interventions specifically designed for youth may provide a cost-effective and engaging alternative to prevent loss of intervention benefits. However, until now online interventions have relied on human moderators to deliver therapeutic content. More sophisticated models responsive to user data are critical to inform tailored online therapy. Thus, integration of user experience with a sophisticated and cutting-edge technology to deliver content is necessary to redefine online interventions in youth mental health. This paper discusses the development of the moderated online social therapy (MOST) web application, which provides an interactive social media-based platform for recovery in mental health. We provide an overview of the system's main features and discus our current work regarding the incorporation of advanced computational and artificial intelligence methods to enhance user engagement and improve the discovery and delivery of therapy content.Methods: Our case study is the ongoing Horyzons site (5-year randomized controlled trial for youth recovering from early psychosis), which is powered by MOST. We outline the motivation underlying the project and the web application's foundational features and interface. We discuss system innovations, including the incorporation of pertinent usage patterns as well as identifying certain limitations of the system. This leads to our current motivations and focus on using computational and artificial intelligence methods to enhance user engagement, and to further improve the system with novel mechanisms for the delivery of therapy content to users. In particular, we cover our usage of natural language analysis and chatbot technologies as strategies to tailor interventions and scale up the system.Conclusions: To date, the innovative MOST system has demonstrated viability in a series of clinical research trials. Given the data-driven opportunities afforded by the software system, observed usage patterns, and the aim to deploy it on a greater scale, an important next step in its evolution is the incorporation of advanced and automated content delivery mechanisms.

Sexual Self-Schemas in the Real World: Investigating the Ecological Validity of Language-Based Markers of Childhood Sexual Abuse.

This is the first study to examine language use and sexual self-schemas in natural language data extracted from posts to a large online forum. Recently, two studies applied advanced text analysis techniques to examine differences in language use and sexual self-schemas between women with and without a history of childhood sexual abuse. The aim of the current study was to test the ecological validity of the differences in language use and sexual self-schema themes that emerged between these two groups of women in the laboratory. Archival natural language data were extracted from a social media website and analyzed using LIWC2015, a computerized text analysis program, and other word counting approaches. The differences in both language use and sexual self-schema themes that manifested in recent laboratory research were replicated and validated in the large online sample. To our knowledge, these results provide the first empirical examination of sexual cognitions as they occur in the real world. These results also suggest that natural language analysis of text extracted from social media sites may be a potentially viable precursor or alternative to laboratory measurement of sexual trauma phenomena, as well as clinical phenomena, more generally.

Center-shared sliding ensemble of neural networks for syntax analysis of natural language

Fixing input sites in ensembles restrict learning movable and distant patterns.The new ensemble changes the sites across component classifiers sharing one position.The ensemble improves accuracy in learning movable patterns and its lower bound.The neural network framework with the ensemble improves accuracy in syntax analysis.The framework is accurate as state-of-the-art methods in Spanish dependency parsing. In this paper, we introduce a new ensemble method specialized to sequential labeling for syntax analysis and propose a neural network framework adopting the ensemble for dependency parsing of natural sentences. The ensemble method assigns sliding input sites to component classifiers which commonly include the position of the label to predict. The method improves labeling accuracy compared to simple ensemble with weighted voting if critical input features have flexible and long distance from the position to predict over sentences. We show the impact of the ensemble through theoretical estimation of its lower bound accuracy and through empirical analysis in a toy problem varying the strength of movability of critical input features. We apply the proposed neural network framework to the two phases of dependency parsing: dependency and relation tagging. Additionally, we newly define the dependency tagging problem using relative dependency and provide a post-processing method to build correct parse trees. In the practical dependency parsing of Spanish IULA corpus, applying the ensemble instead of the simple weighted voting significantly improves accuracy by 0.09% in relation tagging and by 0.06% to 1.59% with respect to the comparison settings in dependency tagging. The framework shows at least 0.28% improvement in the unlabeled attachment score and 0.14% in the labeled attachment score compared to state-of-the-art dependency parsers.

Language, Structure, and Reuse in the Electronic Health Record.

Medical language is at the heart of the electronic health record (EHR), with up to 70 percent of the information in that record being recorded in the natural language, free-text portion. In moving from paper medical records to EHRs, we have opened up opportunities for the reuse of this clinical information through automated search and analysis. Natural language, however, is challenging for computational methods. This paper examines the tension between the nuanced, qualitative nature of medical language and the logical, structured nature of computation as well as the way in which these have interacted with each other through the medium of the EHR. The paper also examines the potential for the computational analysis of natural language to overcome this tension.

Semantic Similarity from Natural Language and Ontology Analysis

Semantic Similarity from Natural Language and Ontology Analysis

Treating conduct disorder: An effectiveness and natural language analysis study of a new family-centred intervention program

This paper reports on a new family-centred, feedback-informed intervention focused on evaluating therapeutic outcomes and language changes across treatment for conduct disorder (CD). The study included 26 youth and families from a larger randomised, controlled trial (Ronan et al., in preparation). Outcome measures reflected family functioning/youth compliance, delinquency, and family goal attainment. First- and last-treatment session audio files were transcribed into more than 286,000 words and evaluated through the Linguistic Inquiry and Word Count Analysis program (Pennebaker et al., 2007). Significant outcomes across family functioning/youth compliance, delinquency, goal attainment and word usage reflected moderate-strong effect sizes. Benchmarking findings also revealed reduced time of treatment delivery compared to a gold standard approach. Linguistic analysis revealed specific language changes across treatment. For caregivers, increased first person, action-oriented, present tense, and assent type words and decreased sadness words were found; for youth, significant reduction in use of leisure words. This study is the first using lexical analyses of natural language to assess change across treatment for conduct disordered youth and families. Such findings provided strong support for program tenets; others, more speculative support.