Natural Language Systems Research Articles

Grounding with Structure: Exploring Design Variations of Grounded Human-AI Collaboration in a Natural Language Interface

Selecting an effective utterance among countless possibilities that match a user's intention poses a challenge when using natural language interfaces. To address the challenge, we leveraged the principle of least collaborative effort in communication grounding theory and designed three grounded conversational interactions: 1) a grounding interface allows users to start with a provisional input and then invite a conversational agent to complete their input, 2) a multiple grounding interface presents multiple inputs for the user to select from, and 3) a structured grounding interface guides users to write inputs in a structure best understood by the system. We compared our three grounding interfaces to an ungrounded control interface in a crowdsourced study (N=80) using a natural language system that generates small programs. We found that the grounding interfaces reduced cognitive load and improved task performance. The structured grounding interface further reduced speaker change costs and improved technology acceptance, without sacrificing the perception of control. We discuss the implications of designing grounded conversational interactions in natural language systems.

Validating Silent Gesture Lab Studies in a Naturally Emerging Sign Language:How Order is Used to Describe Intensional Versus Extensional Events in Nicaraguan Sign Language.

Languages are neither designed in classrooms nor drawn from dictionaries-they are products of human minds and human interactions. However, it is challenging to understand how structure grows in these circumstances because generations of use and transmission shape and reshape the structure of the languages themselves. Laboratory studies on language emergence investigate the origins of language structure by requiring participants, prevented from using their own natural language(s), to create a novel communication system and then transmit it to others.Because the participants in these lab studies are already speakers of a language, it is easy to question the relevance of lab-based findings to the creation of natural language systems. Here, we take the findings from a lab-based language emergence paradigm and test whether the same pattern is also found in a new natural language: Nicaraguan Sign Language. We find evidence that signers of Nicaraguan Sign Language may show the same biases seen in lab-based language emergence studies: (1) they appear to condition word order based on the semantic dimension of intensionality and extensionality, and (2) they adjust this conditioning to satisfy language-internal order constraints. Our study adds to the small, but growing literature testing the relevance of lab-based studies to natural language birth, and provides convincing evidence that the biases seen in the lab play a role in shaping a brand new language.

Beyond the Scalpel: Assessing ChatGPT's Potential as an Auxiliary Intelligent Virtual Assistant in Oral Surgery

AI has revolutionized the way we interact with technology. Noteworthy advances in AI algorithms and large language models (LLM) have led to the development of natural generative language (NGL) systems such as ChatGPT. Although these LLM can simulate human conversations and generate content in real time, they face challenges related to the topicality and accuracy of the information they generate. This study aimed to assess whether ChatGPT-4 could provide accurate and reliable answers to general dentists in the field of oral surgery, and thus explore its potential as an intelligent virtual assistant in clinical decision making in oral surgery. Thirty questions related to oral surgery were posed to ChatGPT4, each question repeated 30 times. Subsequently, a total of 900 responses were obtained. Two surgeons graded the answers according to the guidelines of the Spanish Society of Oral Surgery, using a three-point Likert scale (correct, partially correct/incomplete, and incorrect). Disagreements were arbitrated by an experienced oral surgeon, who provided the final grade Accuracy was found to be 71.7%, and consistency of the experts' grading across iterations, ranged from moderate to almost perfect. ChatGPT-4, with its potential capabilities, will inevitably be integrated into dental disciplines, including oral surgery. In the future, it could be considered as an auxiliary intelligent virtual assistant, though it would never replace oral surgery experts. Proper training and verified information by experts will remain vital to the implementation of the technology. More comprehensive research is needed to ensure the safe and successful application of AI in oral surgery.

An investigation study on the interpretation of ultrasonic medical reports using OpenAI's GPT-3.5-turbo model.

Ultrasound medical reports are an important means of diagnosing diseases and assessing treatment effectiveness. However, their professional terms and complex sentences often make it difficult for ordinary people to understand. Therefore, this study explores the clinical value of using artificial intelligence systems based on ChatGPT to interpret ultrasound medical reports. In this study, a combination of online and offline questionnaires were used to survey both physicians and non-medical individuals. The questionnaires evaluated ChatGPT's interpretation of ultrasound reports from both professional and comprehensibility perspectives, and the results were analyzed using Excel spreadsheets. Additionally, a portion of the research content was evaluated using the Likert Scale 5-point method in the questionnaire. According to survey results, 67.4% of surveyed doctors believe that using ChatGPT for interpreting ultrasound medical reports can help improve work efficiency. At the same time, 69.72% of non-medical professionals believe it is necessary to enhance their understanding of medical ultrasound reports through ChatGPT interpretation, and 62.58% support the application of ChatGPT to ultrasound medical reports. The non-medical group's understanding of ultrasound medical reports significantly improved (p < 0.01) after implementing ChatGPT, However, 67.49% of the general public are concerned about ChatGPT's imperfect functionality, which may cause misleading information. This reflects that the public's trust in new technology is not high enough, and they are also worried about possible privacy leaks and security issues with ChatGPT technology. The higher acceptance and support of non-medical individuals for the interpretation of medical reports by ChatGPT might be due to the system's natural language processing abilities that allow them to better understand and evaluate report contents. However, the expertise and experience of physicians are still irreplaceable. This suggests that the ChatGPT-based ultrasound medical report interpretation system has certain clinical value and application prospects, but further optimization is necessary to address its shortcomings in data quality and professionalism. This study provides a reference and inspiration for promoting the application and development of ultrasound technology and artificial intelligence systems in the medical field.

ChatGPT—Preliminary Overview with Implications for Medicine and Oncology

AbstractThis review provides an overview about the OpenAI system's natural language chat bot called ChatGPT. It focuses on the preliminary assessment of its unique features, advantages, limitations, role in manuscript writing, value in oncology, and future implications.

Movie recommender chatbot based on Dialogflow

&lt;p&gt;&lt;span lang="EN-US"&gt;Currently, the online movie streaming business is growing rapidly, such as Netflix, Disney+, Amazon Prime Video, HBO, and Apple TV. The recommender system helps customers in getting information about movies that are in accordance with their wishes. Meanwhile, the development of messaging platform technology has made it easier for many people to communicate instantly. Utilizing a messaging platform to build a recommender system for movies, provides special benefits because people often access the messaging platform all the time. In the Indonesian language, there are many slang terms that the system must recognize. In this study, we build a chatbot on a messaging platform which users can interact with the system in natural language (in Indonesian language) and get recommendations. We use rule-based and maximum likelihood as a method in natural language processing (NLP), and content-based filtering for the recommendation process. The recommender system interaction is built through a conversation mechanism that will form a conversational recommender system. The interaction is based on a chatbot which is built using Dialogflow and implemented on the telegram. We use the accuracy of recommendations and user satisfaction to evaluate the system performance. The results obtained from the user study indicate that the NLP approach provides a positive experience for users. In addition, the system also produces an accuracy value of 83%.&lt;/span&gt;&lt;/p&gt;

Optimization Logic and Paradigm Construction of Digital Transformation of Technical Skill Training Model under the Perspective of Social Media Communication and Intelligent Natural Language Integration

Research has shown that social media can help students learn. Through a variety of different social platforms, teachers can work with students to integrate social media into their daily instruction to make lessons more interesting and engaging and increase student engagement. With the rise of Metaverse, Artificial Intelligence, Blockchain, New generation communication technology, Brain interface and other intelligent technologies, especially the application of shared media communication and intelligent natural language system, human society has undergone profound changes in production and lifestyle. The digital economy will lead to systematic changes in the concept, mode and governance system of talent training, and the numerical transformation of education training mode, especially the training of technical skills, is imperative. Although some studies try to discuss the construction of intelligent ecology of vocational education, there is a lack of verified scientific theories and practical experience on the implementation path of digital transformation of vocational education. To address this theme, this paper follows the main line of the path of development background, realistic dilemma, hierarchical features, logical framework and paradigm construction of digital transformation of vocational education, and systematically elaborates the great differences between traditional economy and digital economy as the supply base of vocational education development on the basis of full research on the development of digital industry, digital technology and digital education at home and abroad, clarifies the development features including operation mode, strategic direction and value It also focuses on the logical framework of vocational education with industrial chain as the core, innovation chain as the leader, talent chain as the support and education chain as the foundation, analyzes the importance of social communication and natural language intelligent changes in the process of technical skills talents, and explores the practical path of building a vocational education talent training model that is compatible with the intelligent era.

Impact of Artificial Intelligence (AI) For Decision-Making in Organisation

Since the beginning of time, innovation has been the primary force behind rising living standards. But, because innovation renders outdated technology obsolete, it is a very disruptive process. The development of artificial intelligence may have the greatest effects on organisational decision-making of all the new technologies that emerged in the late 20th century. The consequences of its application in complicated social settings have not been adequately studied because the development of artificial intelligence technologies and models has mostly been focused on psychological models of human cognition. This paper aims to generate research that will advance our understanding of artificial intelligence’s effects and function within complex organisations. The linkages between AI technologies and the components of organisational decision-making have been examined in a set of 11 hypotheses that have been created. Here, it is stated that implementing expert systems will result in less complex political decision-making processes, but implementing a natural language system will result in more complex political decision-making processes.

Selecting a set of semantic labels to eliminate ambiguity for Vietnamese

The rapid development of automatic control systems in natural language, automatic translation systems based on semantic statistics have been receiving much attention from computer science researchers. However, this method requires a large bilingual corpus and accurate semantic tagging, the construction of which requires a lot of time and effort, because of the ambiguity of the natural language. For Vietnamese, automatic question-and-answer systems are increasingly developing in Vietnam, but the problem of semantic ambiguity has not yet received much attention from domestic studies. In this paper, we build a model to evaluate and select an effective and reasonable set of semantic labels from 3 commonly used sets of semantic labels: LLOCE (Longman Lexicon of Contemporary English), LDOCE (Longman Dictionary) of Contemporary English) and WordNet. And then, select the appropriate set of labels, apply it to automatic semantic labeling systems for Vietnamese, help eliminate semantic ambiguity, and support automatic translation, automatic question-and-answer systems efficiently.

Reference and morphology*

AbstractThe dominant tradition in analytic philosophy of language views reference as paradigmatically enabled by the acquisition of words from other speakers. Via chains of transmission, these words connect the referrer to the referent. Such a picture assumes the notion of a word as a stable mapping between sound and meaning. Utterances are constructed out of such stable mappings. While this picture of language is both intuitive and historically distinguished, various trends and programs that have developed over the last few decades in theoretical linguistics suggest an alternative. According to these approaches, the word, conceived of as a linguistic ‘building block’, has no special theoretical significance. While natural language systems generate structures mapping sounds onto meanings, they need not do so by composing elements which themselves specify such mappings. I shall describe some of these developments, and show how they pose a problem for traditional philosophical views of language and communication, before identifying an alternative approach to reference which does not rely on this common‐sense picture of words.

Review of the Performance Metrics for Natural Language Systems for Clinical Trials Matching.

Natural Language Processing (NLP) has been adopted widely in clinical trial matching for its ability to process unstructured text that is often found in electronic health records. Despite the rise in the new tools that use NLP to match patients to eligible clinical trials, the comparison of these tools is difficult due to the lack of consistency in how these tools are evaluated. The ground truth or reference that the tools use to assess results varies, making it difficult to compare the robustness of the tools against each other. This paper alarms the lack of definition and consistency of ground truth data used to evaluate such tools and suggests two ways to define a gold standard for the ground truth in small and large-scale studies.

Task-Oriented User Evaluation on Critiquing-Based Recommendation Chatbots

Dialogue-based conversational recommender systems (DCRSs) have become a new trend in recommender systems (RSs), allowing users to communicate with the system in natural language to facilitate feedback provision and product exploration. However, little work has been done to empirically study user perception of and interaction with such systems and, more importantly, how to best support users in providing feedback on the recommendation they receive. In this article, we aim to develop effective <i>critiquing</i> mechanisms for DCRS to improve its feedback elicitation process (i.e., allowing users to <i>critique</i> the current recommendation during the dialogue). Specifically, we have implemented three prototype systems featuring three different critiquing techniques, respectively, i.e., <i>user-initiated critiquing, progressive system-suggested critiquing</i>, and <i>cascading system-suggested critiquing</i>. We have then conducted two task-oriented user studies involving 292 subjects to evaluate the three prototypes. In particular, we consider two typical types of user tasks in RSs: basic recommendation task (BRT, i.e., looking for items according to the user&#x2019;s preferences), and exploration-oriented task (EOT, i.e., exploring different types of items). Results show that EOT stimulates more user interaction, while BRT results in higher user satisfaction. Moreover, when users perform EOT, the type of critiquing techniques is more likely to influence user perception and moderate the relationships between certain interaction metrics and users&#x2019; perceived serendipity. The findings suggest effective critiquing techniques to enhance the interaction between users and the recommendation chatbot when the system makes recommendations for different purposes.

모음체계의 보편성과 유형에 대한 재고

This work aims to revisit universals and typology of vowel systems in natural language. In general, the universals of a vowel system are evaluated by four criteria: the size of the vowel inventory of the language, the nature of constitutional entities of the vowel system, the phonetic features of the vowels, and the distribution of vowels in the vowel space. However, there are languages that are evaluated more correctly when other standards are applied. In this paper, we considered some examples of such languages. Furthermore, in terms of vowel typology, there are vowel systems that cannot be properly approached with a standard view. How to analyze such vowel systems is also discussed. In addition, we discussed the vowel space from a new perspective.

Redefining Multimodality

The term “multimodality” incorporates visible gestures as part of language, a goal first put forward by Adam Kendon, and this idea revolutionized the scope of linguistic inquiry. But here I show that the term “multimodality” itself is rife with ambiguity, sometimes referring to different physical channels of transmission (auditory vs. visual), and sometimes referring to the integration of linguistic structures with more imagistic, less conventionalized expressions (see David McNeill's work), regardless of the physical channel. In sign languages, both modes are conveyed in a single, visual channel, revealed here in the signing of actors in a sign language theatre. In spoken languages, contrary to expectations raised by defining “modality” in terms of the physical channel, we see that the channel of transmission is orthogonal to linguistic and gestural modes of expression: Some visual signals are part and parcel of linguistic structure, while some auditory (intonational) signals have characteristics of the gestural mode. In this empirical, qualitative study, I adopt the term “mode” to refer solely to specific characteristics of communicative expression, and not to the physical channel. “Multimodal” refers to the coexistence of linguistic and gestural modes, regardless of the physical channel of transmission—straightforwardly encompassing the two natural language systems, spoken and signed.

On Information Hiding in Natural Language Systems

With data privacy becoming more of a necessity than a luxury in today's digital world, research on more robust models of privacy preservation and information security is on the rise. In this paper, we take a look at Natural Language Steganography (NLS) methods, which perform information hiding in natural language systems, as a means to achieve data security as well as confidentiality. We summarize primary challenges regarding the secrecy and imperceptibility requirements of these systems and propose potential directions of improvement, specifically targeting steganographic text quality. We believe that this study will act as an appropriate framework to build more resilient models of Natural Language Steganography, working towards instilling security within natural language-based neural models.

Semantic Modeling and Analysis of Natural Language System Requirements

System requirements specify how a system meets stakeholder needs. They are a partial definition of the system under design in natural language that may be restricted in syntax terms. Any natural language specification inevitably lacks a unique interpretation and includes underspecified terms and inconsistencies. If the requirements are not validated early in the system development cycle and refined, as needed, specification flaws may cause costly cycles of corrections in design, implementation and testing. However, validation should be based on a consistent interpretation with respect to a rigorously defined semantic context of the domain of the system. We propose a specification approach that, while sufficiently expressive, it restricts the requirements definition to terms from an ontology with precisely defined concepts and semantic relationships in the domain of the system under design. This enables a series of semantic analyses, which guide the engineer towards improving the requirement specification as well as eliciting tacit knowledge. The problems addressed are prerequisites to enable the derivation of verifiable specifications. We present the results from a case study of modest size from the space system domain, as well as an evaluation of our approach from the user&#x2019;s point of view. The requirement types that have been covered demonstrate the applicability of the approach in an industrial context, although the effectiveness of the analysis depends on pre-existing domain ontologies.

A intersecção entre as Teorias linguísticas e a Linguística Computacional ao longo do tempo

ABSTRACT Recent achievements have turned Computational linguistics into a dynamic research area, and an important field of application development that is being explored by leading technology companies. Despite the advances, there is still much room for improvement to allow humans to interact with computational devices, through natural language, in the same way that humans interact with native speakers of the same language. How to make computers understand or create new metaphors, metonymies or other figures of language? Is it possible to develop natural language systems capable of producing lyrics or poetry on the same level as humans? Can we produce systems capable of assigning new meanings to syntactic elements that are immediately perceived as coherent by humans? In this paper, we account for the evolution of computational linguistics, drawing a parallel with the evolution of linguistic theories and speculating about its research opportunities and foreseeable limitations.

Functional Aspects of Language Games in Political Discourse

The article analyzes various aspects of communication in political discourse. One of its major arguments says that politicians who disrespect major rules of communication violate the basic principles of interaction, namely in that they introduce their own new patterns of language games. Adequate principles of interaction are an indispensable requirement for political discourse as such in that they guarantee efficient communication and help avoid conflicts. Such principles are based on general rules of communication. The article underlines the importance of a structual logical chain of political communiction along the following lines: intention – strategy – tactic – means of realization. In general, human beings acquire languages according to basic linguistic forms and models. Тhe Austrian philosopher L. Wittgenstein introduced the term “language game”. In a similar vein, the Swiss linguist F. de Saussure established a link between language and the sphere of games in that he compared the systems of natural languages to the rules of chess. The language games of political discourse are represented by imperative intentions, emotive senses and various means of the manipulative use of linguistic units; they usually focus on factors of impact. The article discusses the intentions of language games as a point of activization of cognitive and communicative activities for the achievement of goals related to strategies and tactics of politicians. The nature of these intentions exerts impact on the realization of the linguo-pragmatic potential of the functions of perlocutionary linguistic acts (requests, orders, etc.). The perlocutionary functions of language are the basis of political manipulation which manifests itself as a systematic combination of traditional political instruments with contemporary communicative approaches to various aspects of the manipulative impact on political consciousness and behavior. Average citizens who react to political discourse automatically take part in broader intellectual and communivative activities; they also demonstrate their own participation in political processes. Therefore political discourse has an impact on the formation and development of civic society and its relations to the state. As a rule, politicians develop their communicative strategies along programs and platforms that are designed for central subjects of the political process (the government, political parties and leaders). This limited circle of addressees causes a certain lack of efficiency that should be corrected inasmuch as the ultimate goal of any communicative strategy should be a certain change of the addressee’s worldviews. Political tactic is an important component of political communicative technologies. Political communicative activities include various aspects of tactical measures and methods that start on a local level and go viral according to strategically devised lines. In this setting, specific political texts with their particular pragmatic implications appear to be units of political discourse with their own semiotic structure. Politicians create messages based on their world views and ideologies. The role of the addressee should be understood as a priority in political communicative acts that are usually characterized by polemics, a high degree of axiology and persuasiveness. The article focusses on linguistic tools that politicians use in order to manipulate the electorate. To influence the electorate, politicians use manipulative linguistic items on the lexical, idiomatic and metaphorical level. In political discourse, such linguistic items often turn into stereotypical linguistic tools of particular politicians that leave their mark on the electorate’s ideas of their values and beliefs. The article emphasizes the importance of the use of “strong” linguistic items that help create, in the best case, an image of the politician that includes humoristic associations with his or her individual professional language.

A Computational Framework for Slang Generation

AbstractSlang is a common type of informal language, but its flexible nature and paucity of data resources present challenges for existing natural language systems. We take an initial step toward machine generation of slang by developing a framework that models the speaker’s word choice in slang context. Our framework encodes novel slang meaning by relating the conventional and slang senses of a word while incorporating syntactic and contextual knowledge in slang usage. We construct the framework using a combination of probabilistic inference and neural contrastive learning. We perform rigorous evaluations on three slang dictionaries and show that our approach not only outperforms state-of-the-art language models, but also better predicts the historical emergence of slang word usages from 1960s to 2000s. We interpret the proposed models and find that the contrastively learned semantic space is sensitive to the similarities between slang and conventional senses of words. Our work creates opportunities for the automated generation and interpretation of informal language.

MinStab: Stable Network Evolution Rule Mining for System Changeability Analysis

Growing number of evolving systems creates demand for system evolution analytics with modern computational intelligence algorithms and tools. In this paper, we introduce new measures of stability and changeability for system evolution analysis over time. We proposed a Stable Network Evolution Rule Mining and a Changeability Metric for an evolving system. For this, we use two different characteristics of Network Evolution Rules (NERs). First, given a network of a system state S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i</sub> , we call an NER interesting in S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i</sub> if its support and confidence exceed given thresholds (minimum support and minimum confidence). Second, given a set of networks for a set of states (SS), we define the stability of an NER to be the percentage of states in SS in which the rule is interesting. We call an NER stable in SS if its stability exceeds a given threshold named as minimum stability (minStab). Based on this, we developed an intelligent tool, which is used for experiments on evolving systems. We applied our approach to a number of real-world systems including: software system, natural language system, retail market system, and IMDb system. It results Stable NERs and Changeability Metric value for each evolving system.