Conversational Dialogue Systems Research Articles

A Review of AI-Driven Conversational Chatbots Implementation Methodologies and Challenges (1999–2022)

A conversational chatbot or dialogue system is a computer program designed to simulate conversation with human users, especially over the Internet. These chatbots can be integrated into messaging apps, mobile apps, or websites, and are designed to engage in natural language conversations with users. There are also many applications in which chatbots are used for educational support to improve students’ performance during the learning cycle. The recent success of ChatGPT also encourages researchers to explore more possibilities in the field of chatbot applications. One of the main benefits of conversational chatbots is their ability to provide an instant and automated response, which can be leveraged in many application areas. Chatbots can handle a wide range of inquiries and tasks, such as answering frequently asked questions, booking appointments, or making recommendations. Modern conversational chatbots use artificial intelligence (AI) techniques, such as natural language processing (NLP) and artificial neural networks, to understand and respond to users’ input. In this study, we will explore the objectives of why chatbot systems were built and what key methodologies and datasets were leveraged to build a chatbot. Finally, the achievement of the objectives will be discussed, as well as the associated challenges and future chatbot development trends.

Trends and challenges of Arabic Chatbots: Literature review

A conversational system is a natural language processing task that has recently attracted increasing attention with the advancements in Large Language Models (LLMs) and Language Models for Dialogue Applications (LaMDA). However, Conversational Artificial Intelligence (AI) research has mainly been carried out in English. Despite the growing popularity of Arabic as one of the most widely used languages on the Internet, only a few studies have concentrated on Arabic conversational dialogue systems thus far. In this study, we conduct a comprehensive qualitative analysis of the key research works in this domain, examining the limitations and strengths of existing approaches. We start with chatbot history and classification. Then, we examine approaches that leverage Arabic chatbots Rule-based/Retrieval-based and Deep learning-based. In particular, we survey the evolution of Generative Conversational AI with the evolution of deep-learning techniques. Next, we look at the different metrics used to assess conversational systems. Finally, we outline language Challenges for building Generative Arabic Conversational AI.

Reinforcement Learning on Social Chabot to Improve the Semantic Variation

Artificial intelligence (AI) in the fields of conversation and language has experienced immense growth as a result of the large quantities of text corpora available for training models. This paper discusses conversational AI dialogue systems concerning the components of natural language processing (NLP) and reinforcement learning that function together to produce a human-like response. The types of conversational AI, namely, task-oriented systems, question-answering agents, and social chatbots, are also expanded on to provide a systematic review. Reinforcement learning has a crucial role in resolving errors in existing conversational AI models and allows for a factor of exploration in the art of communication. The trial-and-error nature of reinforcement learning agents ensures that irrelevant patterns in a text corpus are not memorized. The implementation of reinforcement learning and NLP tools to create social chatbots with emotional intelligence brings researchers one step closer to mimicking human conversation.

Revisiting the Boundary between ASR and NLU in the Age of Conversational Dialog Systems

Abstract As more users across the world are interacting with dialog agents in their daily life, there is a need for better speech understanding that calls for renewed attention to the dynamics between research in automatic speech recognition (ASR) and natural language understanding (NLU). We briefly review these research areas and lay out the current relationship between them. In light of the observations we make in this article, we argue that (1) NLU should be cognizant of the presence of ASR models being used upstream in a dialog system’s pipeline, (2) ASR should be able to learn from errors found in NLU, (3) there is a need for end-to-end data sets that provide semantic annotations on spoken input, (4) there should be stronger collaboration between ASR and NLU research communities.

Response type selection for chat-like spoken dialog systems based on LSTM and multi-task learning

We propose a method of automatically selecting appropriate responses in conversational spoken dialog systems by explicitly determining the correct response type that is needed first, based on a comparison of the user’s input utterance with many other utterances. Response utterances are then generated based on this response type designation (back channel, changing the topic, expanding the topic, etc.). This allows the generation of more appropriate responses than conventional end-to-end approaches, which only use the user’s input to directly generate response utterances. As a response type selector, we propose an LSTM-based encoder–decoder framework utilizing acoustic and linguistic features extracted from input utterances. In order to extract these features more accurately, we utilize not only input utterances but also response utterances in the training corpus. To do so, multi-task learning using multiple decoders is also investigated.To evaluate our proposed method, we conducted experiments using a corpus of dialogs between elderly people and an interviewer. Our proposed method outperformed conventional methods using either a point-wise classifier based on Support Vector Machines, or a single-task learning LSTM. The best performance was achieved when our two response type selectors (one trained using acoustic features, and the other trained using linguistic features) were combined, and multi-task learning was also performed.

Survey on evaluation methods for dialogue systems

In this paper, we survey the methods and concepts developed for the evaluation of dialogue systems. Evaluation, in and of itself, is a crucial part during the development process. Often, dialogue systems are evaluated by means of human evaluations and questionnaires. However, this tends to be very cost- and time-intensive. Thus, much work has been put into finding methods which allow a reduction in involvement of human labour. In this survey, we present the main concepts and methods. For this, we differentiate between the various classes of dialogue systems (task-oriented, conversational, and question-answering dialogue systems). We cover each class by introducing the main technologies developed for the dialogue systems and then present the evaluation methods regarding that class.

End-to-End Knowledge-Routed Relational Dialogue System for Automatic Diagnosis

Beyond current conversational chatbots or task-oriented dialogue systems that have attracted increasing attention, we move forward to develop a dialogue system for automatic medical diagnosis that converses with patients to collect additional symptoms beyond their self-reports and automatically makes a diagnosis. Besides the challenges for conversational dialogue systems (e.g. topic transition coherency and question understanding), automatic medical diagnosis further poses more critical requirements for the dialogue rationality in the context of medical knowledge and symptom-disease relations. Existing dialogue systems (Madotto, Wu, and Fung 2018; Wei et al. 2018; Li et al. 2017) mostly rely on datadriven learning and cannot be able to encode extra expert knowledge graph. In this work, we propose an End-to-End Knowledge-routed Relational Dialogue System (KR-DS) that seamlessly incorporates rich medical knowledge graph into the topic transition in dialogue management, and makes it cooperative with natural language understanding and natural language generation. A novel Knowledge-routed Deep Q-network (KR-DQN) is introduced to manage topic transitions, which integrates a relational refinement branch for encoding relations among different symptoms and symptomdisease pairs, and a knowledge-routed graph branch for topic decision-making. Extensive experiments on a public medical dialogue dataset show our KR-DS significantly beats stateof-the-art methods (by more than 8% in diagnosis accuracy). We further show the superiority of our KR-DS on a newly collected medical dialogue system dataset, which is more challenging retaining original self-reports and conversational data between patients and doctors.

Dialogue Systems and Conversational Agents for Patients with Dementia: The Human-Robot Interaction.

This study aimed to identify and describe the fundamental characteristics of spoken dialogue systems, and their role in supporting human-robot interaction and enabling the communication between socially assistive robots and patients with dementia. First, this work provides an overview of spoken dialogue systems by considering the underlying technologies, approaches, methods, and general issues. Then, the analysis focuses on studies, systems, and approaches that have investigated the role of dialogue systems and conversational agents in the interaction with elderly people with dementia by presenting the results of a literature review. While the overview of spoken dialogue systems relies on existing surveys and reviews, a research was conducted to identify existing works in the literature that have investigated the role of conversational agents and dialogue systems in the elderly and people with cognitive impairments. Inclusion criteria were as follows: (1) use of conversational agents, dialogue systems, or language processing tools for people with cognitive impairments; (2) age ≥60 years; (3) diagnosis of dementia according to National Institute on Aging-Alzheimer's Association (NIAAA) criteria; (4) presence of tests or experiments with qualitative or quantitative results. Initially 125 studies published between 2000 and 2017 were identified, of which 12 met the inclusion criteria. The review identifies the issues and challenges that are reported when conversational agents and speech-based interfaces have been used for interacting with people with cognitive impairments. In addition, the review led to the identification of studies that have investigated speech processing and natural language processing capabilities to assess the cognitive status of people with dementia.

Interactive Hesitation Synthesis: Modelling and Evaluation

Conversational spoken dialogue systems that interact with the user rather than merely reading the text can be equipped with hesitations to manage dialogue flow and user attention. Based on a series of empirical studies, we elaborated a hesitation synthesis strategy for dialogue systems, which inserts hesitations of a scalable extent wherever needed in the ongoing utterance. Previously, evaluations of hesitation systems have shown that synthesis quality is affected negatively by hesitations, but that they result in improvements of interaction quality. We argue that due to its conversational nature, hesitation synthesis needs interactive evaluation rather than traditional mean opinion score (MOS)-based questionnaires. To validate this claim, we dually evaluate our system’s speech synthesis component, on the one hand, linked to the dialogue system evaluation, and on the other hand, in a traditional MOS way. We are thus able to analyze and discuss differences that arise due to the evaluation methodology. Our results suggest that MOS scales are not sufficient to assess speech synthesis quality, leading to implications for future research that are discussed in this paper. Furthermore, our results indicate that synthetic hesitations are able to increase task performance and that an elaborated hesitation strategy is necessary to avoid likability issues.

Conversational In-Vehicle Dialog Systems: The past, present, and future

Automotive technology rapidly advances with increasing connectivity and automation. These advancements aim to assist safe driving and improve user travel experience. Before the realization of a full automation, in-vehicle dialog systems may reduce the driver distraction from many services available through connectivity. Even when a full automation is realized, in-vehicle dialog systems still play a special role in assisting vehicle occupants to perform various tasks. On the other hand, in-vehicle use cases need to address very different user conditions, environments, and industry requirements than other uses. This makes the development of effective and efficient in-vehicle dialog systems challenging; it requires multidisciplinary expertise in automatic speech recognition, spoken language understanding, dialog management (DM), natural language generation, and application management, as well as field system and safety testing. In this article, we review research and development (R&D) activities for in-vehicle dialog systems from both academic and industrial perspectives, examine findings, discuss key challenges, and share our visions for voice-enabled interaction and intelligent assistance for smart vehicles over the next decade.

Building End-To-End Dialogue Systems Using Generative Hierarchical Neural Network Models

We investigate the task of building open domain, conversational dialogue systems based on large dialogue corpora using generative models. Generative models produce system responses that are autonomously generated word-by-word, opening up the possibility for realistic, flexible interactions. In support of this goal, we extend the recently proposed hierarchical recurrent encoder-decoder neural network to the dialogue domain, and demonstrate that this model is competitive with state-of-the-art neural language models and back-off n-gram models. We investigate the limitations of this and similar approaches, and show how its performance can be improved by bootstrapping the learning from a larger question-answer pair corpus and from pretrained word embeddings.

Introduction to the special section on eye gaze and conversation

This editorial introduction first explains the origin of this special section. It then outlines how each of the two articles included sheds light on possibilities for conversational dialog systems to use eye gaze as a signal that reflects aspects of participation in the dialog: degree of engagement and turn taking behavior, respectively.

MushyPeek: A Framework for Online Investigation of Audiovisual Dialogue Phenomena

Evaluation of methods and techniques for conversational and multimodal spoken dialogue systems is complex, as is gathering data for the modeling and tuning of such techniques. This article describes MushyPeek, an experiment framework that allows us to manipulate the audiovisual behavior of interlocutors in a setting similar to face-to-face human-human dialogue. The setup connects two subjects to each other over a Voice over Internet Protocol (VoIP) telephone connection and simultaneously provides each of them with an avatar representing the other. We present a first experiment which inaugurates, exemplifies, and validates the framework. The experiment corroborates earlier findings on the use of gaze and head pose gestures in turn-taking.

Multi-pass recognition of spoken dialogue

A system and method of multi-pass recognition for conversational spoken dialogue systems includes two speech recognizers: a first recognizer that implements, for example, a statistical language model (SLM) and a second recognizer that implements, for example, a grammar-based model. A word-spotting speech recognizer may be included, as may confidence estimators for each speech recognizer. The system and method provide a multi-pass approach to speech recognition, which reevaluates speech inputs to improve recognition where confidence scores returned from confidence estimators are low.

Computing the fundamental frequency variation spectrum in conversational spoken dialogue systems

Continuous modeling of intonation in natural speech has long been hampered by a focus on modeling pitch, of which several normative aspects are particularly problematic. The latter include, among others, the fact that pitch is undefined in unvoiced segments, that its absolute magnitude is speaker-specific, and that its robust estimation and modeling, at a particular point in time, rely on a patchwork of long-time stability heuristics. In the present work, we continue our analysis of the fundamental frequency variation (FFV) spectrum, a recently proposed instantaneous, continuous, vector-valued representation of pitch variation, which is obtained by comparing the harmonic structure of the frequency magnitude spectra of the left and right half of an analysis frame. We analyze the sensitivity of a task-specific error rate in a conversational spoken dialogue system to the specific definition of the left and right halves of a frame, resulting in operational recommendations regarding the framing policy and window shape.

Let's Chat: A conversational dialogue system for second language practice

Early and intermediate second language (L2) learners often encounter difficulties when engaging in introductory social conversations, typically having had little opportunity to practise such interactions. This article describes a project to design and prototype a computer dialogue system, Let's Chat, which would allow learners to rehearse social conversations without a human partner. While most existing systems do not facilitate interactions of a sufficiently high quality to be useful for L2 acquisition, Let's Chat does so by relying on communication through pre-stored phrases, thereby steering clear of the technological limitations of conventional natural language processing. This approach is supported by research that suggests that the holistic assimilation of L2 sequences can boost the communicative capabilities and fluency of learners. By practising with Let's Chat, learners would acquire good target language input in the form of well-constructed, native-like utterances and thus be better-prepared, more confident participants during natural conversations.

Modeling Non-Verbal Behavior in Multimodal Conversational Systems (Modellierung nicht-verbalen Verhaltens in Multimodalen Dialogsystemen)

Summary Non-verbal behavior is an integral part of conversational dialog. When people engage in dialogue they utilize non-verbal behavior both to structure the flow of the conversation as well as to provide feedback about their current understanding of the discourse while the other one is speaking. In this article, we investigate the function of non-verbal behavior and show how it is integrated into a discourse model used within a conversational dialogue system.

Trainable approaches to surface natural language generation and their application to conversational dialog systems

We study how decisions for word ordering and word choice in surface natural language generation can be automatically learned from annotated data. We examine four trainable systems for surface natural language generation in the air travel domain, called NLG[1–4]. NLG1 is a lookup table which stores the most frequent phrase to express a concept, and is intended as a baseline system for comparison purposes. NLG2 and NLG3 attempt to find the highest probability word sequence with respect to a maximum entropy probability model. They differ in that NLG2 predicts words left-to-right, while NLG3 predicts words in dependency tree order. NLG4 requires a dependency-style grammar of phrase fragments and conditions on their use, and attempts to find the highest probability word sequence that is consistent with the rules and conditions of the grammar. NLG4 has been implemented in a dialog strategy for a prototype air travel conversational system, in which word order is dynamically modified to emphasize certain aspects of the run-time dialog state.