Dialogue Model Research Articles

MedChatZH: A tuning LLM for traditional Chinese medicine consultations

Generative Large Language Models (LLMs) have achieved significant success in various natural language processing tasks, including Question-Answering (QA) and dialogue systems. However, most models are trained on English data and lack strong generalization in providing answers in Chinese. This limitation is especially evident in specialized domains like traditional Chinese medical QA, where performance suffers due to the absence of fine-tuning and high-quality datasets. To address this, we introduce MedChatZH, a dialogue model optimized for Chinese medical QA based on transformer decoder with LLaMA architecture. Continued pre-training on a curated corpus of Chinese medical books is followed by fine-tuning with a carefully selected medical instruction dataset, resulting in MedChatZH outperforming several Chinese dialogue baselines on a real-world medical dialogue dataset. Our model, code, and dataset are publicly available on GitHub (https://github.com/tyang816/MedChatZH) to encourage further research in traditional Chinese medicine and LLMs.

Science communication objectives and actual practices of science news websites as a showcase for gaps between theory and practice

This study contributes to the growing body of science communication research showing gaps between theory and practice objectives, focusing on one particular understudied and emerging science communication innovation.The objectives and practices of four Israeli science news websites were analyzed considering three science communication models: “Dissemination”, “Dialogue”, and “Participation.” Using concurrent parallel mixed methods, we examined the perspectives of website administrators (n=8) and readers (n=20) through interviews, a content analysis of news items (n=298), discussion threads (n=507), and reader questionnaires (n=89). Findings indicate limited adoption of two-way communication about how science is applied in society. The scant implementation of the dialogue model suggests its promises are not concretized in practice on these science news websites.

Dialogue as a Counterextremism Policy: Lessons from the History of Early Islam—The Kharijites vs. the Authority

precis: Governments worldwide have made it a priority to combat religious extremism, resulting in various disciplinary and soft-power policies. This study focuses on dialogue as an effective way to counter and prevent religious extremism, using the Kharijites (the first extremist faction that mutinied against the caliphal authority) as a case study. Historical methodology is used to collect data, and a critical analysis is conducted using inductive and deductive methods to draw conclusions and insights for a proposed dialogue model. The essay argues that military power is ephemeral and sometimes counterproductive, so the emphasis should be on a long-lived "mind power." Further, given the embedded nature of religious underpinnings of extremists' ideologies, the study highlights seven key terms that form the foundation of the three-level dialogue model.

Distinct but correct: generating diversified and entity-revised medical response

Medical Dialogue Generation (MDG) is intended to build a medical dialogue system for intelligent consultation, which can communicate with patients in real-time, thereby improving the efficiency of clinical diagnosis with broad application prospects. This paper presents our proposed framework for the Chinese MDG organized by the 2021 China conference on knowledge graph and semantic computing (CCKS) competition, which requires generating context-consistent and medically meaningful responses conditioned on the dialogue history. In our framework, we propose a pipeline system composed of entity prediction and entity-aware dialogue generation, by adding predicted entities to the dialogue model with a fusion mechanism, thereby utilizing information from different sources. At the decoding stage, we propose a new decoding mechanism named Entity-revised Diverse Beam Search (EDBS) to improve entity correctness and promote the length and quality of the final response. The proposed method wins both the CCKS and the International Conference on Learning Representations (ICLR) 2021 Workshop Machine Learning for Preventing and Combating Pandemics (MLPCP) Track 1 Entity-aware MED competitions, which demonstrate the practicality and effectiveness of our method.

Enhancing multi-modal fusion in visual dialog via sample debiasing and feature interaction

Visual dialog aims to accomplish multiple rounds of dialog by fusing information extracted from images, captions, and previous question–answer pairs. As a vision-language task, visual dialog encounters challenges related to language bias and vision bias. These biases create an imbalance in multi-modal fusion, resulting in shortcut learning and significantly compromising the model’s robustness. Moreover, existing multi-modal fusion methods in visual dialog exhibit a low data interaction frequency, leading to insufficient fusion. To overcome the balance and sufficiency issues in multi-modal fusion, we propose a novel Parallel Attention Fusion visual dialog model with Counterfactual Sample debiasing (CS-PAF). Specifically, CS-PAF consists of two core ingredients: (i) a counterfactual sample generation module for model debiasing; and (ii) a parallel attention fusion network that enhances sufficiency in multi-modal data interaction. Notably, in contrast to other debiasing methods, our counterfactual sample generation applies contrastive learning to circumvent the high cost of manual annotations and ensure seamless integration with other models. Extensive comparisons with state-of-the-art approaches, along with comprehensive ablation and transferability studies across multiple datasets, substantiate the superiority and effectiveness of our CS-PAF. Our implement code is available at https://github.com/chenyulu2000/cspaf.

Feature Extraction of Dialogue Text Based on Big Data and Machine Learning

In this article, a dialogue text feature extraction model based on big data and machine learning is constructed, which transforms the high-dimensional space of text features into the low-dimensional space that is easy to process, so that the best feature words can be selected to represent the document set. Tests show that in most cases, the classification accuracy of this model is higher than 88%, and the recall rate is higher than 85%, thus achieving the goal of higher classification accuracy with less computation. When extracting the features of dialogue texts, there is no need for preprocessing, just count the data such as lexical composition, sentence length and sentence-to-sentence relationship of the target text, and make linear analysis to obtain key indicators and weights. Based on this, the classification model can achieve good results, thus effectively reducing the workload and computation of text classification.

Memutus Mata Rantai Stunting: Transformasi Pendampingan Keluarga Oleh Tim Stunting Kecamatan Kunduran Blora

This research aims to formulate a transformation model for stunting assistance by the Kunduran District Stunting Team, Blora Regency. The study uses qualitative methods with a phenomenological approach through in-depth interviews. The results of the research show that the implementation of the role of companion for stunting families in Kunduran District still faces several obstacles, such as limited competence of companions, less effective counselling methods, barriers to access to health services for stunting families, and early detection and supplementation of at-risk pregnant women. A transformation of the stunting assistance paradigm is needed from the old top-down approach to a dialogical and collaborative model between the assistant and the family. Transformation is required by increasing companions' competence, interactive counselling, an organized referral system, early detection of at-risk pregnant women, and increasing access to nutritional supplementation. With comprehensive transformation, it is hoped that supporting stunted families will be more effective in accelerating the reduction of stunting in Kunduran District and Blora Regency. Regional government support through policies and funding is urgently needed to optimize the transformation of the role of family companions.

A Mood Semantic Awareness Model for Emotional Interactive Robots.

The rapid development of natural language processing technology and improvements in computer performance in recent years have resulted in the wide-scale development and adoption of human-machine dialogue systems. In this study, the Icc_dialogue model is proposed to enhance the semantic awareness of moods for emotional interactive robots. Equipped with a voice interaction module, emotion calculation is conducted based on model responses, and rules for calculating users' degree of interest are formulated. By evaluating the degree of interest, the system can determine whether it should transition to a new topic to maintain the user's interest. This model can also address issues such as overly purposeful responses and rigid emotional expressions in generated replies. Simultaneously, this study explores topic continuation after answering a question, the construction of dialogue rounds, keyword counting, and the creation of a target text similarity matrix for each text in the dialogue dataset. The matrix is normalized, weights are assigned, and the final text score is calculated. In the text with the highest score, the content of dialogue continuation is determined by calculating a subsequent sentence with the highest similarity. This resolves the issue in which the conversational bot fails to continue dialogue on a topic after answering a question, instead waiting for the user to voluntarily provide more information, resulting in topic interruption. As described in the experimental section, both automatic and manual evaluations were conducted to validate the significant improvement in the mood semantic awareness model's performance in terms of dialogue quality and user experience.

Engaging young people in science communication about mental health during COVID-19

Many young people struggle with their mental health and the COVID-19 pandemic compounded these challenges. However, young people are rarely involved in research and communication about causes and coping strategies. We used an online game as a conversation starter and co-created a list of coping strategies with young people to apply the dialogue model of science communication and facilitate social conversation about mental health during COVID-19. The young people found the involvement was valuable as it led to self-reflection, social reflection with peers and an experience of recognition and contribution. We discuss challenges and urge researchers to explore ways for open dialogue and co-creation as strategic and contributing parts of the research process.

TodBR: Target-Oriented Dialog with Bidirectional Reasoning on Knowledge Graph

Target-oriented dialog explores how a dialog agent connects two topics cooperatively and coherently, which aims to generate a “bridging” utterance connecting the new topic to the previous conversation turn. The central focus of this task entails multi-hop reasoning on a knowledge graph (KG) to achieve the desired target. However, current target-oriented dialog approaches suffer from inefficiencies in reasoning and the inability to locate pertinent key information without bidirectional reason. To address these limitations, we present a bidirectional reasoning model for target-oriented dialog implemented on a commonsense knowledge graph. Furthermore, we introduce an automated technique for constructing dialog subgraphs, which aids in acquiring multi-hop reasoning capabilities. Our experiments demonstrate that our proposed method attains superior performance in reaching the target while providing more coherent responses.

Общество в поисках безопасной модели межкультурного диалога

The article is devoted to the problem of dialogue of cultures in the aspect of its security, ensuring the preservation of their integrity and survival. In the course of analyzing the history of European civilization, the authors identified three models of dialogue between cultures in the context of ensuring safe existence: 1. The ancient “separation” model; 2. Classical unification model of the New Age; 3. Modern open dialogue model. The main attention is paid to the analysis of the open model of dialogue of cultures, which is proposed as a model of security. The theoretical source of the research is primarily Russian philosophy of the second half of the 20th century, in particular the philosophy of dialogue by M. M. Bakhtin and V. S. Bibler.

Construction and Evaluation of a Machine Learning-Based Conversation Model for English Spoken Tourism Conversation

Abstract The development of language macromodeling has led to the widespread adoption of spoken English conversation systems in various industries. Similarly, it has had an impact on the tourism industry, where the involvement of machine learning has taken a big step forward. The purpose of this paper is to explore the use of machine learning methods in spoken dialog systems to construct a dialog model that is suitable for the tourism industry. Speech, map image, and tourism information text data are collected first, which form the database of the dialog model through processing, storage, and application. A bag-of-words model and word vectors represent the text, while speech activity and an English conversation system construct the speech and intention recognition model. On the test set, the best speech recognition rate using double-threshold endpoint detection is up to 98.73%, and the corresponding sentence recognition rate is also up to 93.67%. This paper’s model achieves the highest intention recognition accuracy, scoring 0.943 and 0.955 on the CCKS2018 and SMP2018 datasets, respectively. This paper proposes a machine-learning conversation model that provides technical support for conversational AI development in the tourism industry.

GPT Modeling for English Learning Assistance Functions in Non-Native Language Environments

Abstract In this paper, we design an English multi-round dialogue model based on the pre-trained model GPT-2 and use Top-p and Top-k sampling instead of greedy search and cluster search sampling to optimize the decoding strategy. After that, a dialog history keyword replication mechanism is introduced to improve context consistency. Based on the GPT-2 English Multi-Round Dialogue Model, a conversational teaching model for teaching English in a non-native language environment has been constructed. The impact of this teaching model on students’ theoretical learning and teaching skills was examined. The results show that the BLEU1 index value of the GPT-2DH model after pre-training in the chatterbot corpus is 0.356, which is almost twice as much as that of seq2seq. As a result, the model in this paper is superior. Compared with traditional teaching methods, the GPT dialogic learning model significantly improved students’ theoretical learning performance, English listening and speaking skills, learning attitude (P=0.029<0.05), and extrinsic learning motivation (P=0.046<0.05) but had no significant effect on their instructional design skills and intrinsic learning motivation.

Towards Building an Open-Domain Dialogue System Incorporated With Internet Memes

In recent years, Internet memes have been widely used in online chatting. Compared with text-based communication, conversations become more expressive and attractive when Internet memes are incorporated. This paper presents our solutions for the Meme incorporated Open-domain Dialogue (MOD) Challenge of DSTC10, where three tasks are involved: text response modeling, meme retrieval, and meme emotion classification. Firstly, we leverage a large-scale pre-trained dialogue model for coherent and informative response generation. Secondly, based on interaction-based text-matching, our approach can retrieve appropriate memes with good generalization ability. Thirdly, we propose to model the emotion flow (EF) in conversations and introduce an auxiliary task of emotion description prediction (EDP) to boost the performance of meme emotion classification. Experimental results on the MOD dataset demonstrate that our methods can incorporate Internet memes into dialogue systems effectively.

Classroom Utterance Analysis and Visualization Using a Generative Deep Neural Networks for Dialogue Model

Classroom Utterance Analysis and Visualization Using a Generative Deep Neural Networks for Dialogue Model

Global cultural dialogue model at the cultural meetings of the 21st century Azerbaijani intellectuals

Global cultural dialogue model at the cultural meetings of the 21st century Azerbaijani intellectuals

The Interfunctional Relationship Between Theory of Mind and Private Speech.

Theory of mind is the capacity to explain and predict the behavior of others. Charles Fernyhough's dialogical model of psychological functions offers a vision of theory of mind that considers the social dimension and the importance of language, especially inner and private speech, for a person's ability to represent and manipulate multiple perspectives, and its connection to executive function. There is little direct research on Fernyhough's model. The present study addressed that gap by studying the relationship between theory of mind, private speech, and executive function (planning) in the preschool years. Data were collected from a total of 86 preschool children from the Mexican state of Tlaxcala; 24 were from the first grade of preschool, 30 from the second, and 32 from the third. Their degree of internalization and their speech types (i.e., social and private) were obtained by coding their utterances during free play and during performance of a Tower of London task. Lastly, their theory of mind was assessed with a change of location and an unexpected contents task. No correlation was found between children's theory of mind and their degree of internalization. However, inaudible private speech was correlated (negatively) to theory of mind performance in the third grade of preschool. Furthermore, their scores on the Tower of London task were negatively correlated with social speech and positively correlated with inaudible private speech, while the reverse was true for time of execution. The results suggested that the degree of internalization is a measure of the regulative function of language, not the ability to manipulate perspectives, and that it is inadequate for capturing subtle differences between performance and speech types. Role-playing conditions were recommended as better tasks for exploring the capacities for manipulating and understanding different perspectives during the preschool years.

Interreligious Environmentalism: The Way Ahmadiyya Group Engages to Interfaith Dialogue

There have been many studies on Ahmadiyya and their involvement in inter/intra-religious dialogue. However, some of these studies only look at how the dialogue takes shape or tends to be formal. Ahmadiyya, which is a vulnerable religious group, certainly needs a new model of dialogue or encounter. This article assumes that "mutual collaboration in public issues" can be an encounter space for interfaith dialogue for religious groups, especially for minorities. From this, this research aims to see the strategy of the Ahmadiyya community in Yogyakarta to be able to engage with society and also see the inter/intra-religious dialogue model used by the Ahmadiyya community in Yogyakarta. To answer these questions, this research uses qualitative research methods. Furthermore, this research takes data through the interview process and literature obtained from Ahmadiyya online media and then analyzed using descriptive and interpretive analysis techniques. This study found that the Ahmadiyya community uses environmental issues to have dialogue with other religious communities and become a space for encounter and acceptance in society.

초등학생의 공감코칭프로그램 개발 및 효과

The purpose of this study is to develop a coaching program to improve elementary school students' empathy for peer relationship attachment and a happy and enjoyable school life. Elementary students can motivate themselves by using coaching techniques to present their opinions rather than the vertical teachings of teachers, and to use the C-PLUS coaching dialogue model to give positive feedback to elementary students. Rather than giving answers
 on how to improve abilities, various activities helped elementary school students to internalize and live empathy. To examine the effect of the empathy coaching program for elementary school students developed in this study, self-regulation and communication skills were divided into pre-test and post-test. The even higher scores of self-regulation and communication in the experimental group when compared to the control group suggest that the cognitive, emotional, and communicative elements of the empathy coaching program were properly arranged. In addition, the program goal of the researcher was consistent in that elementary students' empathy was improved by improving their empathy ability by strengthening their positive emotion and internal control through coaching conversation. This study demonstrates the need for coaching programs and the value of their use.

Dialog Antaragama Dan Kontribusi Tokoh Agama Dalam Penyelesaian Konflik Dan Implementasinya Untuk Memperkuat Toleransi

In a peaceful society, it is a common need that cannot be avoided amidst differences. The process of dialogue between religious communities becomes important when we talk about the relationship between each religion and belief. Because without dialogue, the ideals of peace and harmony become just empty shadows. This is also due to the existence of dialogue model activities, which is one form of realizing peaceful life in a multicultural society. These results also prove that the existence of an inter-religious dialogue model in conflict resolution is a means of strengthening tolerance between religious groups carried out by society. also as a model for dialogue in community life that meets each other and carries out dialogue on social issues. Apart from that, dialogue on religious experiences, which involves all elements of society, themselves carry out faith traditions, namely prayer based on the guidance of their respective faiths in joint prayer activities at the village hall every year. Religious experience activities are carried out represented by each religious figure. to lead the prayers which are performed alternately. Therefore, the dialogue model between religious communities can be concluded that the life dialogue model of conflict resolution that strengthens tolerance in a pluralistic society can build religious experience and action dialogue that develops in particular communities and this is a very effective solution in building a peaceful community life in particular. in order to create harmony among different people in the midst of our pluralistic society.