China
Africa
Explore chapters and articles related to this topic
Topic-Specific Natural Language Chatbot as General Advisor for College
Published in Nilanjan Dey, Sanjeev Wagh, Parikshit N. Mahalle, Mohd. Shafi Pathan, Applied Machine Learning for Smart Data Analysis, 2019
Varun Patil, Yogeshwar Chaudhari, Harsh Rohila, Pranav Bhosale, P. S. Desai
We have proposed to develop a system which gives a precise and to-the-point answers to the queries of a user rather than making user search for answers amongst a huge pool of data. The system has an access to all topic-specific data and answers some personal info queries. As for natural language processing, we use Dialogflow, formerly known as Api.ai, a service provided by Google, which helps us to set entities and its actions in the form of intents which helps us to make a chatbot which is domain specific. Thus, this domain-specific chatbot acts as an advisor to the users in a particular domain. We have incorporated an optimized context awareness to increase efficiency of the chatbot. We have included a voice-based system–user interaction that increases the usability of this system. The user asks queries in the form of voice and system gives a voice-based answer. This could thus help physically challenged people as well. We have explored other events related to the domain and found an easy way to implement it in one go. For example, consider a college conducting a survey where a notification could be given by the June that a new survey has been added to the system for a particular group of students and has to be completed. It could be then completed upon the user’s approval. It will replicate the user experience as one in a personal interview. A provision can be added to provide results for non-domain-related/unanswered queries by looking into other data sources such as Google search results link, Wikipedia, and so on
Visual design of dialogue flows for conversational interfaces
Published in Behaviour & Information Technology, 2021
Stefano Valtolina, Lorenzo Neri
DialogFlow is a natural language processing (NLP) platform that can be used to build conversational applications and experiences on multiple platforms (e.g. Facebook, Messenger) or devices (e.g. Google Home). Developers are supported by tools to enhance their app’s interaction features through AI-powered text and voice discussions. DialogFlow supports effectively the designer in implementing the communicative strategies described in Table 2. To do it, DialogFlow offers a design procedure that allows the designer to focus on different parts of the creative process that concern both the design of the flow and the coding of the functionalities necessary to create the bot's replies. To design the flow, the designer has to specify basic intents that contain the following components: Training phrases: samples of the phrases that the users can say.Action and Parameters: training phrases can be annotated with entities or categories of data that the developer wishes DialogFlow to match. They have the purpose of improving the intent's language model.Responses: samples of the text, speech, or visual responses to provide to the users, which usually prompts users in a way that lets them know what to say next or that the conversation is ending up.
Influence of Rapport and Social Presence with an AI Psychotherapy Chatbot on Users’ Self-Disclosure
Published in International Journal of Human–Computer Interaction, 2022
Jieon Lee, Daeho Lee, Jae-gil Lee
Socially-aware chatbots have been proposed by various researchers (e.g., Jain et al., 2018). The counseling chatbot for this study is a fully automated text-based program implemented using Google Dialogflow.1 Dialogflow is a development platform that enables a chatbot to give plausible responses to a wide range of user inputs based on natural language processing technology, providing them with a natural conversation experience (M. Lee et al., 2019; S. Lee et al., 2020; Santoso et al., 2018).
Towards a decision support system for radiotherapy business continuity in a pandemic crisis
Published in Journal of Decision Systems, 2022
Melanie Reuter-Oppermann, Ralf Müller-Polyzou, Anthimos Georgiadis
In the first design cycle, the DSS system has been instantiated as a conversational AI based chatbot. We used Google Dialogflow to model the natural user interface and to realise the knowledge base (Google, 2020). The chatbot is implemented as a web service to provide support on any device with a web browser. The user communicates with the RT-Companion agent that consists of a welcome intent, the knowledge base, an acquisition, a BCM and RM as well as fallback intents. Every intent is based on training phases, actions, parameters and responses as shown in Figure 5. The wel- come intent gives an introduction to the topic and provides an overview of the risk mitigation themes. The knowledge base consists of 97 separate intents reflecting the identified risk mitigation measures. The acquisition intent offers users the possibility to share best practices and experience. The BCM and RM intent provides information about BCM and RM in the RT context. The welcome and the BCM and RM intents are implemented with follow-up intents based on a yes-no user dialogue. The NLP module interprets the user’s language and translates text or audio to structured data. Dialogflow uses two algorithms for the intent classification. Rule-based grammar matching and machine learning (ML) matching work simultaneously, then the best result is chosen. The ML algorithm of the agent is trained to match conversations to intents. During this process, the user input is compared to intent training phrases to find the best matching intent. The intent matching was monitored in a test phase and manually corrected with the support of the Dialogflow built-in validation module. In particular, the 97 intents of the knowledge base were trained individually to ensure a clear assignment, especially for adjacent mitigation measures. The flow chart of the RT-Companion agent is outlined in Figure 6. The user interface of the conversational agent is shown in Figure 7. RT-Companion in its version 1.0 is capable of providing information about risk mitigation themes, BCM and RM. It can collect new mitigation measures provided by users. Most important, it provides users with answers within the scope of the risk mitigation knowledge base. The agent also supports text-to-speech and can thereby read out the information provided.