Conversational Language Research Articles

Explaining machine learning models with interactive natural language conversations using TalkToModel

Practitioners increasingly use machine learning (ML) models, yet models have become more complex and harder to understand. To understand complex models, researchers have proposed techniques to explain model predictions. However, practitioners struggle to use explainability methods because they do not know which explanation to choose and how to interpret the explanation. Here we address the challenge of using explainability methods by proposing TalkToModel: an interactive dialogue system that explains ML models through natural language conversations. TalkToModel consists of three components: an adaptive dialogue engine that interprets natural language and generates meaningful responses; an execution component that constructs the explanations used in the conversation; and a conversational interface. In real-world evaluations, 73% of healthcare workers agreed they would use TalkToModel over existing systems for understanding a disease prediction model, and 85% of ML professionals agreed TalkToModel was easier to use, demonstrating that TalkToModel is highly effective for model explainability.

Eos and OMOCL: Towards a seamless integration of openEHR records into the OMOP Common Data Model

Background:The reuse of data from electronic health records (EHRs) for research purposes promises to improve the data foundation for clinical trials and may even support to enable them. Nevertheless, EHRs are characterized by both, heterogeneous structure and semantics. To standardize this data for research, the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) standard has recently seen an increase in use. However, the conversion of these EHRs into the OMOP CDM requires complex and resource intensive Extract Transform and Load (ETL) processes. This hampers the reuse of clinical data for research. To solve the issues of heterogeneity of EHRs and the lack of semantic precision on the care site, the openEHR standard has recently seen wider adoption. A standardized process to integrate openEHR records into the CDM potentially lowers the barriers of making EHRs accessible for research. Yet, a comprehensive approach about the integration of openEHR records into the OMOP CDM has not yet been made. Methods:We analyzed both standards and compared their models to identify possible mappings. Based on this, we defined the necessary processes to transform openEHR records into CDM tables. We also discuss the limitation of openEHR with its unspecific demographics model and propose two possible solutions. Results:We developed the OMOP Conversion Language (OMOCL) which enabled us to define a declarative openEHR archetype-to-CDM mapping language. Using OMOCL, it was possible to define a set of mappings. As a proof-of-concept, we implemented the Eos tool, which uses the OMOCL-files to successfully automatize the ETL from real-world and sample EHRs into the OMOP CDM. Discussion:Both Eos and OMOCL provide a way to define generic mappings for an integration of openEHR records into OMOP. Thus, it represents a significant step towards achieving interoperability between the clinical and the research data domains. However, the transformation of openEHR data into the less expressive OMOP CDM leads to a loss of semantics.

Analisis Alih Kode dalam Novel “Senja & Pagi” Karya Alffy Rev

This study analyzes code switching and code mixing in the novel "Senja&Pagi" by Alffy Rev. The aim of research on code-switching in languages is to understand and analyze the phenomenon of using more than one language or variations of languages in certain conversations or communication contexts. There is an external form of code switching in the form of language switching which includes: a) Indonesian to Arabic switching and b) Indonesian to English switching, and there is also internal code switching, namely c) Indonesian to Javanese switching. In the novel "Senja & Pagi" by Alffy Rev there are factors that cause code switching due to the presence of a speaker or speakers, interlocutors or listeners, changes in subject matter and topics of conversation.

Tingkat Tutur pada Cerkak di Web Page Jawasastra

Javanese is one of the unique regional languages ​​because it has speech levels which are shown through spoken language in daily life conversations, while forms of written language are usually applied through Javanese literary works. This article aims to analyze the form and factors that cause the occurrence of speech levels used by the characters in cêrkak that represent people's lives. This study was studied using a qualitative descriptive method and analyzed using a sociolinguistic approach in the speech level study. The research data is sourced from documents of speech level that originating from dialogues between characters in five cêrkak on the Jawasastra web page. Research subjects were selected through purposive sampling, while data was collected through document analysis. The data were analyzed with an interactive analysis model through data collection, data reduction, data presentation, and drawing conclusions. In this study, four variations of speech levels were found, including: 12 data of ngoko lugu, 3 data of ngoko alus, 1 data of krama lugu, and 9 data of krama alus. The factors behind the use of each level of speech within the scope of Javanese society are due to differences in levels of education, economy, age, position, familiarity, and heredity.

An Empirical Study of Adoption of ChatGPT for Bug Fixing among Professional Developers

ChatGPT is a powerful tool that assists software engineers in identifying and rectifying errors in code. One of its primary advantages is its ability to engage in natural language conversation with humans, which allows it to collaborate more closely with engineers in improving and optimizing the code. However, despite its potential advantages, software developers do not always utilize ChatGPT as a tool for bug fixing. In this study, we aim to examine the factors that influence the adoption of ChatGPT for bug fixing among professional software developers, based on the Unified Theory of Acceptance and Use of Technology (UTAUT) theory. To accomplish this, we conducted 50 semi-structured interviews with professional software developers and other stakeholders. Our findings indicate that the performance expectancy and effort expectancy of professional software developers, as well as social influence, facilitating conditions, data security, and trust are the key factors of adoption. These findings suggest that understanding these factors can be critical in promoting the adoption and use of ChatGPT in the software development industry.

Parametric Evaluation Techniques for Reliability of Internet of Things (IoT)

Reliability analysis aims to assess the dependability and performance of a system over time, considering factors such as device failures, communication issues, and data integrity. By integrating taxonomy and reliability analysis, researchers can systematically analyze and enhance the reliability of IoT systems by addressing specific components and categories identified in the taxonomy. IoT based on its reliable applications is classified among various categories such as smart home devices, smart watches, electric vehicles with IoT capabilities, industrial, healthcare, agriculture, soil monitoring sensors, smart city infrastructure, environmental monitoring devices, and smart energy management devices. The need for a consistent taxonomy was one of the problems recognized and examined in this paper even though IoT management technology is developing. To examine the reliability analysis of IoT systems, we created a brand-new taxonomy for IoT gadgets that places a strong focus on administration. We have released a completely new IoT reliability RBD model analysis at different gateways and Round-Trip Time is to be analyzed. It features an aircraft for dependability. Our goal is to help administrators of networks, as well as designers, put processes in place that will analyze and enhance the reliability of IoT systems. To implement techniques to increase the IoT system's reliability, we may utilize the management capabilities to keep an eye on the devices. ChatGPT can act as the conversational interface for ISABELA. By integrating ChatGPT with ISABELA's functionality, users can have natural language conversations with ISABELA, making the interaction more intuitive and user-friendly. A use case study of ISABELA has enhanced the position of the CHATGPT system in the IoT system which increase the efficiency and reliability of IoT designs and networks.

LANGUAGE STYLE USED IN OLYMPUS HAS FALLEN

Abstract
 Language style is used by humans to convey meaning in different depending on their situation of speaking. People even mix their language styles based on context. It makes every human unique since it creates different language style because of their personality and situation in communication. The researcherss found it interesting to analyze the language style used in literary work. Therefore, this research discussed about the types of language style used in movie script, especially Olympus Has Fallen. Qualitative method is applied to conduct this study. The researcherss collected the data by u s i n g non-participant observation and audio-visuals material. The researcherssgot the data by analyzing the script. After that, the researcherss marked theutterances used by the actors and actresses. The researcherss used theory of Keraf (2009) to classify the types of language style. They are formal style, informal style and conversational style. After doing this research, the researcherss found all the types of language styles that mentioned by Keraf (2009). The researcherss found 102 data that were categorized into types of language style that used in movie script of Olympus Has Fallen. The results were classified into 12 data of formal style with 12 data, informal style with 25 data, and conversational with 65 data. The findings were dominated by conversational language style because the conversation is indeed used in daily context by using jargon, idioms, abbreviation, and slang. Overall, the researchers hope this research can add insight in the field of linguistics.

COLA: Improving Conversational Recommender Systems by Collaborative Augmentation

Conversational recommender systems (CRS) aim to employ natural language conversations to suggest suitable products to users. Understanding user preferences for prospective items and learning efficient item representations are crucial for CRS. Despite various attempts, earlier studies mostly learned item representations based on individual conversations, ignoring item popularity embodied among all others. Besides, they still need support in efficiently capturing user preferences since the information reflected in a single conversation is limited. Inspired by collaborative filtering, we propose a collaborative augmentation (COLA) method to simultaneously improve both item representation learning and user preference modeling to address these issues. We construct an interactive user-item graph from all conversations, which augments item representations with user-aware information, i.e., item popularity. To improve user preference modeling, we retrieve similar conversations from the training corpus, where the involved items and attributes that reflect the user's potential interests are used to augment the user representation through gate control. Extensive experiments on two benchmark datasets demonstrate the effectiveness of our method. Our code and data are available at https://github.com/DongdingLin/COLA.

Keeping Active with Texting after Stroke (KATS): development of a text message intervention to promote physical activity and exercise after stroke

BackgroundPost-stroke physical activity reduces disability and risk of further stroke. When stroke rehabilitation ends, some people feel abandoned by services and struggle to undertake physical activities that support recovery and health. The aim of this study was to codesign a novel text message intervention to promote physical activity among people with stroke and provide support when formal rehabilitation ends. This manuscript describes the intervention development processes that will inform future pilot and feasibility studies.MethodsThe planned intervention was a series of text messages to be sent in a predetermined sequence to people with stroke at the end of rehabilitation. The intervention, underpinned by behaviour change theory and using salient behaviour change techniques, would provide daily messages offering encouragement and support for the uptake and maintenance of physical activity following stroke. The intervention was codesigned by a Collaborative Working Group, comprised of people with stroke, rehabilitation therapists, representatives from stroke charities and academics. A four-step framework was used to design the intervention: formative research on physical activity post-stroke, creation of the behaviour change text message intervention, pre-testing and refinement. Formative research included a review of the scientific evidence and interviews with community-dwelling people with stroke. Data generated were used by the Collaborative Working Group to identify topics to be addressed in the intervention. These were mapped to constructs of the Health Action Process Approach, and salient behaviour change techniques to deliver the intervention were identified. The intervention was rendered into a series of text messages to be delivered over 12 weeks. The draft intervention was revised and refined through an iterative process including review by people with stroke, their spouses, rehabilitation therapists and experts in the field of stroke.The messages encourage regular physical activity but do not prescribe exercise or provide reminders to exercise at specific times. They use conversational language to encourage engagement, and some are personalised for participants. Quotes from people with stroke provide encouragement and support and model key behaviour change techniques such as goal setting and coping planning.DiscussionCo-design processes were critical in systematically developing this theory and evidence-based intervention. People with stroke and rehabilitation therapists provided insights into perceived barriers post-rehabilitation and identified strategies to overcome them. The structured multistep approach highlighted areas for improvement through successive rounds of review. The intervention will be tested for acceptability, feasibility and effectiveness in future studies. This co-design approach could be used for interventions for other heath behaviours and with different populations.

Use of Gesture Recognition for Differently Abled Persons

There are nine billion dumb people in the world. The communication between dumb people and normal people is always a challenging task. One of the key technologies to afford reality is usual interaction between human and objects or living-things in the implicit world created in a computer. Disabled persons are an important part of our society. With the advent of science and technology, efforts are being made to develop certain systems that make them feel and act normally. Especially hearing impaired people interact through hand gestures or signs. It is difficult to find a well experienced translator for the sign language conversion every time and everywhere.

What a discourse unit can teach us about the relationship between the Mishnah and Tosefta: an examination of the parallelism between the compilations in the halakhic give-and-take conversation

Abstract This article presents an examination of the parallelism between the Mishnah and Tosefta in one discourse unit – the halakhic give-and-take conversation. It aims to show that a description of discourse units found in both compilations can contribute to the discussion of the relationship between the two compilations and the status of the Tosefta in regard to Mishnah. In the examined corpus of halakhic give-and-take conversations from the Mishnah and Tosefta from three orders, for only 16 conversations in the Tosefta (14%) was there found a parallel conversation in the Mishnah, and in most cases the parallels are not identical. The structural and linguistic comparison between these 16 parallel conversations showed that the conversations in the Tosefta contain more exchanges as well as more complete exchanges, and that the language in the Tosefta seems less redacted and earlier compared with the language in the parallel conversations in the Mishnah.

Analyzing Language in the Picnic Scene Picture and in Conversation: The Type of Discourse Sample We Choose Influences Findings in People With Aphasia.

People with aphasia express that improved conversational discourse is a primary rehabilitation goal. Discourse is usually assessed using monologue, such as a picture description task, but research shows that language in monologue varies from language in everyday conversation. Consequently, we investigated the relationship of language in unstructured conversation and in the picnic scene picture because it is a part of the most often used aphasia battery (Western Aphasia Battery-Revised) and thus is frequently used to inform therapy. Second, because previous research suggests people with severe aphasia may not demonstrate language production variability between types of monologue-level discourse, we evaluated the relationship of severity and the difference in scores between conversation and the picnic scene task. Thirty-four people with mild-to-severe aphasia described the picnic scene and provided a conversation sample. We measured language production and communicative success using seven measures with established psychometrics in conversation/monologue. We conducted correlations to answer the research questions. Correlations were moderate and weaker for the measures in the two conditions. A strong negative relationship was demonstrated between aphasia severity and global coherence. All other relationships were moderate and weaker for the remaining measures when correlated with aphasia severity (also negative). Results are consistent with other studies indicating that language varies in different types of discourse. We conclude that for accurate, meaningful assessment, discourse sampling needs to include the specific type of discourse the individual wishes to address in therapy, because discourse samples and their findings are not interchangeable.

ChatGPT and Open-AI Models: A Preliminary Review

According to numerous reports, ChatGPT represents a significant breakthrough in the field of artificial intelligence. ChatGPT is a pre-trained AI model designed to engage in natural language conversations, utilizing sophisticated techniques from Natural Language Processing (NLP), Supervised Learning, and Reinforcement Learning to comprehend and generate text comparable to human-generated text. This article provides an overview of the training process and fundamental functionality of ChatGPT, accompanied by a preliminary review of the relevant literature. Notably, this article presents the first comprehensive literature review of this technology at the time of publication, aiming to aggregate all the available pertinent articles to facilitate further developments in the field. Ultimately, the authors aim to offer an appraisal of the technology’s potential implications on existing knowledge and technology, along with potential challenges that must be addressed.

Vocabulary skills of bilingual children: The evidence for context‐dependent performance

AbstractMost bilingual children often display greater proficiency and preference for one of their two languages. Researchers refer to this asymmetry as language dominance. However, despite being possible to determine the overall language dominance in bilinguals, there may be substantial flexibility in their language use. In particular, the relative ease of using each language may depend on the sociolinguistic context and the language of the interlocutor which may affect the ease of accessing words in a given language. In the present work, we investigate whether and how the sociolinguistic context in an International English medium school influences bilingual children's vocabulary production. Specifically, we examine whether bilingual first‐grade children's expressive vocabulary skills vary as a function of the linguistic context (Home/School) and the language (English/Spanish) of their conversational partner. Results showed that contextual factors interacted with child's language performance. The participants' ability to name objects in Spanish was affected by the conversational partner's language, and their ability to name Home‐context words varied by language. The findings of this investigation highlight the importance of considering sociolinguistic factors when evaluating bilingual children's language skills.

Relation of Chatbot Usage Towards Customer Satisfaction Level in Indonesia

In customer service for businesses in Indonesia, chatbot are predicted to be a pillar of business, especially chatbot that are able to reach customers proactively, provide availability 24 hours a day, even fulfill requests and answer customer questions quickly which has an important value in customer satisfaction factor in use of chatbot in Indonesia. The world of chatbots is very diverse to study from various aspects such as from a technological point of view, the study of natural language conversation of chatbots done by Shawar and Atwell concluded that chatbots are useful as a “tool”, entertainment tools, also learning and practice tools, as well as tools to help with e-commerce and other fields. Another chatbot study compared the ability of offline customer service agents in general on marketing and how to represent an online business with a chatbot and its effect on improving relationships with customers. In conclusion, chatbots provide an efficient alternative solution to conventional customer service in reducing the repetitive tasks of human admins in solving customer problems, or allowing the ability to respond to various customer needs.

Long-Term Effects of Perceived Friendship with Intelligent Voice Assistants on Usage Behavior, User Experience, and Social Perceptions

Social patterns and roles can develop when users talk to intelligent voice assistants (IVAs) daily. The current study investigates whether users assign different roles to devices and how this affects their usage behavior, user experience, and social perceptions. Since social roles take time to establish, we equipped 106 participants with Alexa or Google assistants and some smart home devices and observed their interactions for nine months. We analyzed diverse subjective (questionnaire) and objective data (interaction data). By combining social science and data science analyses, we identified two distinct clusters—users who assigned a friendship role to IVAs over time and users who did not. Interestingly, these clusters exhibited significant differences in their usage behavior, user experience, and social perceptions of the devices. For example, participants who assigned a role to IVAs attributed more friendship to them used them more frequently, reported more enjoyment during interactions, and perceived more empathy for IVAs. In addition, these users had distinct personal requirements, for example, they reported more loneliness. This study provides valuable insights into the role-specific effects and consequences of voice assistants. Recent developments in conversational language models such as ChatGPT suggest that the findings of this study could make an important contribution to the design of dialogic human–AI interactions.

Explaining Recommendations through Conversations: Dialog Model and the Effects of Interface Type and Degree of Interactivity

Explaining system-generated recommendations based on user reviews can foster users’ understanding and assessment of the recommended items and the recommender system (RS) as a whole. While up to now explanations have mostly been static, shown in a single presentation unit, some interactive explanatory approaches have emerged in explainable artificial intelligence (XAI), making it easier for users to examine system decisions and to explore arguments according to their information needs. However, little is known about how interactive interfaces should be conceptualized and designed to meet the explanatory aims of transparency, effectiveness, and trust in RS. Thus, we investigate the potential of interactive, conversational explanations in review-based RS and propose an explanation approach inspired by dialog models and formal argument structures. In particular, we investigate users’ perception of two different interface types for presenting explanations, a graphical user interface (GUI)-based dialog consisting of a sequence of explanatory steps, and a chatbot-like natural-language interface. Since providing explanations by means of natural language conversation is a novel approach, there is a lack of understanding how users would formulate their questions with a corresponding lack of datasets. We thus propose an intent model for explanatory queries and describe the development of ConvEx-DS, a dataset containing intent annotations of 1,806 user questions in the domain of hotels, that can be used to to train intent detection methods as part of the development of conversational agents for explainable RS. We validate the model by measuring user-perceived helpfulness of answers given based on the implemented intent detection. Finally, we report on a user study investigating users’ evaluation of the two types of interactive explanations proposed (GUI and chatbot), and to test the effect of varying degrees of interactivity that result in greater or lesser access to explanatory information. By using Structural Equation Modeling, we reveal details on the relationships between the perceived quality of an explanation and the explanatory objectives of transparency, trust, and effectiveness. Our results show that providing interactive options for scrutinizing explanatory arguments has a significant positive influence on the evaluation by users (compared to low interactive alternatives). Results also suggest that user characteristics such as decision-making style may have a significant influence on the evaluation of different types of interactive explanation interfaces.

Walking My Path Together with W. W. Phelps

Walking My Path Together with W. W. Phelps

Sign Language Recognition Based on Hands Symbol's Classification

One of the most natural and ancient types of conversational language is sign language. The technology that converts sign language into writing for people who have difficulty communicating, such as those who have speech issues, hearing disabilities, or are deaf, is the subject of this study. This paper is based on a real-time method based on finger writing and neural networks for American sign language. An interesting field of vision study is the automatic recognition of human gestures from video images. The Research recommend employing a convolution neural network (CNN) method to recognize human hand motions from a photograph. The objective is to identify hand movements used in human work activities from a camera image. Hand placement and orientation are employed to collect the training and assessment data for CNN. The hand is first put through a filter, and once that has been done, it is put through a classification, which determines what class the hand movements belong to. Then, CNN is trained using the measured pictures.

ソフトウェア構造情報と動作ログを活用した企業情報システムの実装言語刷新支援方式

Legacy enterprise information systems are required to convert their implementation languages into modern languages to enhance their maintainability. However, in a planning phase, it is difficult for engineers to identify units of programs to be converted appropriately and quickly. In this paper, we propose a support method that can effectively identify the units of programs for language conversion and test data acquisition points by relating application logs to software structure. We implemented this method by using graph database, and through the application of this method to real enterprise information system, we confirmed that we can effectively identify units of program conversion and test data acquisition points.