DOCSLIB.ORG

Learning Representations of Chatbots

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Learning Representations of Chatbots Bot2Vec: Learning Representations of Chatbots Jonathan Herzig1 Tommy Sandbank∗∗;2 Michal Shmueli-Scheuer1 David Konopnicki1 John Richards1 1IBM Research 2Similari fhjon,shmueli,davidkog.il.ibm.com, [email protected], [email protected] Abstract complex objects that execute logic in order to drive conversations with users. How should they best be Chatbots (i.e., bots) are becoming widely used represented? in multiple domains, along with supporting bot programming platforms. These platforms are Many commercial companies provide bot pro- equipped with novel testing tools aimed at im- gramming platforms. These platforms provide proving the quality of individual chatbots. Do- tools and services to develop bots, monitor and ing so requires an understanding of what sort improve their quality. Due to the increasing pop- of bots are being built (captured by their under- ularity of bots, thousands or tens of thousands of lying conversation graphs) and how well they bots could be deployed by different companies on perform (derived through analysis of conver- each platform1. Although bots might have differ- sation logs). In this paper, we propose a new ent purposes and different underlying structures, model, BOT2VEC, that embeds bots to a com- pact representation based on their structure the ability to understand bot behavior at a high and usage logs. Then, we utilize BOT2VEC level could inspire new tools and services bene- representations to improve the quality of two fitting all bots on the platform. In this work, we bot analysis tasks. Using conversation data explore a commercial platform, and study differ- and graphs of over than 90 bots, we show that ent bot representations. BOT2VEC representations improve detection Inspired by the success of recently proposed performance by more than 16% for both tasks. learned embeddings for objects such as graphs 1 Introduction (Narayanan et al., 2017), nodes (Grover and Leskovec, 2016), documents (Le and Mikolov, As conversational systems (i.e., chatbots) become 2014) and words (Mikolov et al., 2013a), we pro- more pervasive, careful analysis of their capabil- pose a new model, BOT2VEC, that learns bot ities becomes important. Conversational systems embeddings, and propose both content and graph are being used for a variety of support, service, based representations. While previous graph em- and sales applications that were formerly handled bedding representations consider static local struc- by human agents. Thus, organizations deploying tures in the graph (Narayanan et al., 2017; Grover such systems must be able to understand bots be- and Leskovec, 2016), our graph representation is havior to improve their performance. In many based on dynamic conversation paths. As bots are cases, such an analysis can be viewed as a classi- usually represented on bot platforms as some form fication task whose goal is to check whether a bot of directed graph, with conversations represented or a particular instance of a conversation satisfies as traversals on the graph, this approach seems some property (e.g., is the conversation success- reasonable. It captures the way the bot is actually ful?). Models for these downstream classification used, in addition to how it is structured. tasks should benefit from conditioning on repre- In this paper, our goal is to consider various bot sentations that capture global bot behavior. embeddings and two different but realistic classi- For a conversation itself, there exists a natural fication tasks, and test whether some bot represen- way to represent it as the concatenation of the hu- tations are more appropriate for these tasks. The man and bot utterances. As for a bot, the question first task, at the level of entire bots, aims to detect of its representation is more complicated: bots are 1https://www.techemergence.com/chatbot-comparison- ∗ Work was done while working at IBM. facebook-microsoft-amazon-google/ 75 Proceedings of the Eighth Joint Conference on Lexical and Computational Semantics (*SEM), pages 75–84 Minneapolis, June 6–7, 2019. c 2019 Association for Computational Linguistics (1) (1,1) (1,1,1) (1,1,1,1) (1,1,1,1,1) Welcome Make_a Payment_ Payment_ Paypal Paypal_error _payment time methods the context of a conversation. (1,2) (1,1,1,2) (1,1,1,1,2) Payment_type Credit_card Insert_detail At every step (i.e., every turn) of a conversation (2) (2,1) (2,1,1) with a bot, the human user expresses an utterance Account_ New_account Create_user operation (2,2) (2,2,1) and the bot analyzes it, determines how to respond Delete_ Delete_user account and updates its internal state. This determination (2,3) Update_ is executed by traversing the graph, starting from account (3) (3,1) a special node called the root node, and moving Store_ Location information (4) (4,1) (4,1,1) (4,1,1,1) along the nodes of the graph according to a given Technical_ Headset_ Wired_ Echo_test problem problem model set of rules as described below. Note that this de- (4,2) (4,1,2) (4,1,2,1) HD_problem Wireless_ Charge_ model headset scription aims to present and explain key abstrac- (4,3) Battery_ Goto: tions rather than the implementation details of an problem Echo_test (5) actual bot platform. Device_ jump node information negative edge (6) positive edge Anything_Else 2.1 Graph Components Figure 1: Example of a customer support bot graph. Every node in the graph has two internal parts: a user intent, and an optional reply of the bot. Given a user utterance, an intent classifier is used to de- termine whether the user utterance matches the in- whether the bot is in production (i.e., in use with tent associated with the node. For example, the real human users) or not. The second task, at the Technical problem node has been defined to cap- conversation level within each bot, aims to detect ture cases where users encounter a technical prob- problematic conversations with a deployed bot in lem with a product, and this is what is being ex- support of focusing improvement efforts. pressed in the utterance at hand (e.g., “I’m hav- The main contributions of this paper are three- ing some issues with my headset”). In this case, fold: (1) this is the first research that leverages in- the classifier should be able to classify this utter- formation from multiple bots in order to improve ance as relevant to this intent. In practice, the in- bot quality, (2) this is the first research to propose tent classifier is trained from examples of utter- an embedding approach for bots based on their ances and their corresponding intents, written by structure and how this structure is exploited dur- bot programmers. ing conversations, and (3) we empirically evaluate Every node has two optional outgoing edges: a these representations for two classification tasks positive edge and a negative edge. If a user ut- using data from more than 90 conversational bots. terance has been classified positively, the optional We find that our proposed representations lead node answer is presented to the user, and the exe- to more than 16% improvement in classification cution moves to the node which is the target of the for our two tasks, with our structure-based repre- positive edge. When the execution moves along sentation performing better than our content-based a positive edge, it is possible to obtain additional representation. This suggests that the representa- input from the user, and continue the evaluation tions explored in this paper are valuable across a using this input. If a user utterance has been clas- range of possible tasks. sified negatively, the execution moves to the node which is the target of the negative edge. As shown 2 Bot Overview in Figure1, we represent nodes connected by neg- Although different programming models can be ative edges along the vertical axis, and nodes con- used to create bots, in practice, most commercial nected by positive edges along the horizontal axis. conversational system platforms represent the con- When there is no positive edge and the classi- versation control flow for bots as graphs. In this fication is positive, the execution stops, and the paper, we use a bot paradigm based on one of the system waits for the next user utterance. When publicly available commercial platforms, but it is there is no negative edge and the classification is quite general and can be adapted to fit to other bot negative, execution jumps back to the root node to programming platforms. Our example in Figure1 resume evaluation from the start. shows a part of a customer support bot graph, and In addition to the nodes that we just described, we use it to explain how such a graph is used in there are special sink nodes (in our example, the 76 Agent: Hi, I’m Eva a virtual assistant, how may I help you? using?”), the system then waits for the next user Customer: I’m having some issues with my headset utterance. The next user utterance (“A wireless Agent: Which model are you using? Customer: A wireless one. one.”) leads to the Wireless model node, hence the Agent: Please make sure that they are charged. resulting path for this utterance is a continuation of Customer: Yes they are, and they work well with other de- the previous path. vices... Agent: Got you. In this case go and try our echo test. Note that nodes connected vertically by neg- Customer: Well, I’ll stop by one of your stores and see a tech- ative edges represent alternative understand- nician. Agent: I’m not trained on that yet, but I’m still learning. You ings of an utterance. That is, in our ex- may want to rephrase your question and try again. ample, an utterance can be identified as Ac- Customer: What is your closest location? count operation, Store information or Techni- Agent: Check here to see our locations in your city.
Load more

Recommended publications
  • The Chatbot Revolution
    The chatbot revolution Moving beyond the hype and maximizing customer experience Ready for a digital 2% 2017 conversation? 25% 2020 Consumers have higher expectations than ever when it comes to interacting with brands. By 2020, 25% of customer They demand personalized service, easily accessible support options, a quick response after reaching out, and successful resolutions on a tight turnaround. service operations will use To meet these needs, companies are increasing their use of digital channels to chatbot or virtual assistant communicate with customers – in fact, by 2022, 70% of all customer interactions will involve technology like messaging applications, social platforms, or chatbots. technologies, an increase Let’s take a closer look at chatbots. Their functions range from answering simple from just 2% in 2017. questions like informing customers of store hours or location to more advanced ones, like handling a credit card charge dispute. According to Gartner, by 2020, 25% of customer service operations will use chatbot or virtual assistant technologies, an increase from just 2% in 2017. When trying to balance staffing budgets, round- the-clock service availability and a preference for digital platforms, chatbots on paper seem like the obvious – and inevitable – choice to engage customers through automation. But how inevitable is it really? 1. Gartner Magic Quadrant for Customer Engagement Center, Michael Maoz, Brian Manusama, 16 May 2018 www.pega.com The chatbot revolution 01 Why the digital hold up? Consumers and businesses have concerns. Despite Gartner predictions and the obvious excitement around chatbots, overall adoption has been slow. Currently most chatbots are programmed to follow predetermined conversational flows—thus limiting their usefulness for solving complex problems or picking up conversational subtleties.
    [Show full text]
  • Voice Interfaces
    VIEW POINT VOICE INTERFACES Abstract A voice-user interface (VUI) makes human interaction with computers possible through a voice/speech platform in order to initiate an automated service or process. This Point of View explores the reasons behind the rise of voice interface, key challenges enterprises face in voice interface adoption and the solution to these. Are We Ready for Voice Interfaces? Let’s get talking! IO showed the new promise of voice offering integrations with their Voice interfaces. Assistants. Since Apple integration with Siri, voice interfaces has significantly Almost all the big players (Google, Apple, As per industry forecasts, over the next progressed. Echo and Google Home Microsoft) have ‘office productivity’ decade, 8 out of every 10 people in the have demonstrated that we do not need applications that are being adopted by world will own a device (a smartphone or a user interface to talk to computers businesses (Microsoft and their Office some kind of assistant) which will support and have opened-up a new channel for Suite already have a big advantage here, voice based conversations in native communication. Recent demos of voice but things like Google Docs and Keynote language. Just imagine the impact! based personal assistance at Google are sneaking in), they have also started Voice Assistant Market USD~7.8 Billion CAGR ~39% Market Size (USD Billion) 2016 2017 2018 2019 2020 2021 2022 2023 The Sudden Interest in Voice Interfaces Although voice technology/assistants Voice Recognition Accuracy Convenience – speaking vs typing have been around in some shape or form Voice Recognition accuracy continues to Humans can speak 150 words per minute for many years, the relentless growth of improve as we now have the capability to vs the typing speed of 40 words per low-cost computational power—and train the models using neural networks minute.
    [Show full text]
  • Intellibot: a Domain-Specific Chatbot for the Insurance Industry
    IntelliBot: A Domain-specific Chatbot for the Insurance Industry MOHAMMAD NURUZZAMAN A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy UNSW Canberra at Australia Defence Force Academy (ADFA) School of Business 20 October 2020 ORIGINALITY STATEMENT ‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institute, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project’s design and conception or in style, presentation and linguistic expression is acknowledged.’ Signed Date To my beloved parents Acknowledgement Writing a thesis is a great process to review not only my academic work but also the journey I took as a PhD student. I have spent four lovely years at UNSW Canberra in the Australian Defence Force Academy (ADFA). Throughout my journey in graduate school, I have been fortunate to come across so many brilliant researchers and genuine friends. It is the people who I met shaped who I am today. This thesis would not have been possible without them. My gratitude goes out to all of them.
    [Show full text]
  • The Inner Circle Guide to AI, Chatbots & Machine Learning
    The Inner Circle Guide to AI, Chatbots & Machine Learning Sponsored by The Inner Circle Guide to AI, Chatbots and Machine Learning © ContactBabel 2019 Please note that all information is believed correct at the time of publication, but ContactBabel does not accept responsibility for any action arising from errors or omissions within the report, links to external websites or other third-party content. 2 Understand the customer experience with the power of AI Employees Customers Businesses Increase agent Elevate customer Gain improved engagement experience VoC insights Artificial Customer Machine Intelligence surveys learning Recorded CRM VoC calls notes analytics Social media Chatbots Surveys opentext.com/explore CONTENTS Contents ..................................................................................................................................................... 4 Table of Figures ........................................................................................................................................... 6 About the Inner Circle Guides ..................................................................................................................... 7 AI: Definitions and Terminology ................................................................................................................. 9 Definitions............................................................................................................................................. 11 Use Cases for AI in the Contact Centre ....................................................................................................
    [Show full text]
  • A Survey on Different Algorithms Used in Chatbot
    International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 05 | May 2020 www.irjet.net p-ISSN: 2395-0072 A survey on Different Algorithms used in Chatbot Siddhi Pardeshi1, Suyasha Ovhal2, Pranali Shinde3, Manasi Bansode4, Anandkumar Birajdar5 1Siddhi Pardeshi, Student, Dept. of Computer Engineering, Pimpri Chinchwad College of Engineering, Pune Maharashtra, India 2Suyasha Ovhal, Student, Dept. of Computer Engineering, Pimpri Chinchwad College of Engineering, Pune Maharashtra, India 3Pranali Shinde, Student, Dept. of Computer Engineering, Pimpri Chinchwad College of Engineering, Pune Maharashtra, India 4Manasi Bansode, Student, Dept. of Computer Engineering, Pimpri Chinchwad College of Engineering, Pune Maharashtra, India 5Professor, Dept. of Computer Engineering, Pimpri Chinchwad College of Engineering, Pune, Maharashtra, India ---------------------------------------------------------------------***---------------------------------------------------------------------- Abstract - Machines are working similar to humans Rule-based/Command based: In these types of chatbots, because of advanced technological concepts. Best example is predefined rules are stored which includes questions and chatbot which depends on advanced concepts in computer answers. Based on what question has requested by the user science. Chatbots serve as a medium for the communication chatbot searches for an answer. But this gives limitations on between human and machine. There are a number of chatbots the type of questions and answers to be stored. and design techniques available in market that perform Intelligent Chatbots/AI Chatbots: To overcome the issue different function and can be implemented in sectors like faced by rule based chatbots intelligent chatbots are business sector, medical sector, farming etc. The technology developed. As these are based on advanced machine learning used for the advancement of conversational agent is natural concepts, they have the ability to learn on their own language processing (NLP).
    [Show full text]
  • MULTILINGUAL CHATBOT with HUMAN CONVERSATIONAL ABILITY [1] Aradhana Bisht, [2] Gopan Doshi, [3] Bhavna Arora, [4] Suvarna Pansambal [1][2] Student, Dept
    International Journal of Future Generation Communication and Networking Vol. 13, No. 1s, (2020), pp. 138- 146 MULTILINGUAL CHATBOT WITH HUMAN CONVERSATIONAL ABILITY [1] Aradhana Bisht, [2] Gopan Doshi, [3] Bhavna Arora, [4] Suvarna Pansambal [1][2] Student, Dept. of Computer Engineering,[3][4] Asst. Prof., Dept. of Computer Engineering, Atharva College of Engineering, Mumbai, India Abstract Chatbots - The chatbot technology has become very fascinating to people around the globe because of its ability to communicate with humans. They respond to the user query and are sometimes capable of executing sundry tasks. Its implementation is easier because of wide availability of development platforms and language libraries. Most of the chatbots support English language only and very few have the skill to communicate in multiple languages. In this paper we are proposing an idea to build a chatbot that can communicate in as many languages as google translator supports and also the chatbot will be capable of doing humanly conversation. This can be done by using various technologies such as Natural Language Processing (NLP) techniques, Sequence To Sequence Modeling with encoder decoder architecture[12]. We aim to build a chatbot which will be like virtual assistant and will have the ability to have conversations more like human to human rather than human to bot and will also be able to communicate in multiple languages. Keywords: Chatbot, Multilingual, Communication, Human Conversational, Virtual agent, NLP, GNMT. 1. Introduction A chatbot is a virtual agent for conversation, which is capable of answering user queries in the form of text or speech. In other words, a chatbot is a software application/program that can chat with a user on any topic[5].
    [Show full text]
  • A Chatbot System Demonstrating Intelligent Behaviour Using
    International Journal of Advanced Research in Computer Engineering & Technology (IJARCET) Volume 4 Issue 10, October 2015 A chatbot system demonstrating Intelligent Behaviour using NLP 1Ameya Vichare, 2Ankur Gyani, 3Yashika Shrikhande, 4Nilesh Rathod 1Student, IT Department, RGIT, Mumbai 2Student, IT Department, RGIT, Mumbai 3Student, IT Department, RGIT, Mumbai 4Assistant Professor, IT Department, RGIT, Mumbai Abstract— Chatbots are computer programs that interact making capabilities, availability of corpora, processing tool with users using natural languages. Just as people use language standards like XML [1]. for human communication, chatbots use natural language to communicate with human users. In this paper, we begin by In our project we are using AIML. AIML is an XML- introducing chatbots and emphasize their need in day-to-day based language which can be used for interaction between activities. Then we go on to discuss existing chatbot systems, chatbots and humans. The atomic unit in AIML is category, namely ELIZA, ALICE and Siri. We evaluate what we can take which consists of attributes called as pattern and template. from each of the systems and use in our proposed system. We are also using a speech to text/text to speech recognition Finally, we discuss the proposed system. The system we intend to develop can be used for educating the user about sports. The to recognize the Indian accent more efficiently. While database will be fed with sports related data which will be constructing the soundbase of the chatbot, the following can coded using AIML. This system can be used in operating help heighten its speech recognition rate: the soundbase systems in a similar manner to Siri for effective information should be built to match user speech input based on retrieval just by speaking various queries.
    [Show full text]
  • Implementation of a Chatbot Using Natural Language Processing Niranjan Dandekar1, Suyog Ghodey2 1,2Department of Computer, Pimpri Chinchwad College of Engineering
    Implementation of a Chatbot using Natural Language Processing Niranjan Dandekar1, Suyog Ghodey2 1,2Department Of Computer, Pimpri Chinchwad College of Engineering ABSTRACT Over the past few years, Artificial Intelligence has grown and leaps and bounds. We are heading towards a society in which machines will take care of the most complex issues that we are confronting today. Before long, the greater part the assignments on the planet will be computerized. This will prompt a huge increment in connection amongst people and machines. This collaboration will be interceded by Natural Language handling. A chatterbot or chatbot intends to make a discussion between both human and machine. The machine has been implanted the learning to distinguish the sentences and choosing itself as reaction to answer a question. The reaction rule is coordinating the info sentence from client. The information of chatbot is put away in the database. Clearly, there is an expansion in the request of talk mechanization on the grounds that a) it expels the human element and b) it can give a 24-hour benefit which will multiplicatively affect the income era. Keywords: Chatbot, Database, Natural Language Processing, Response principle I. INTRODUCTION The development of Artificial Intelligence applications is challenging because computers traditionally require humans to speak to them in a programming language that is precise, unambiguous and highly structured or, perhaps through a limited number of clearly-stated voice commands. Natural language processing(NLP) is a branch of artificial intelligence, and machine linguistics that enables computers to derive meaning from human or natural language input. It is used to analyze text, allowing machines to understand human‟s language.
    [Show full text]
  • Evaluating Natural Language Understanding Services for Conversational Question Answering Systems
    Evaluating Natural Language Understanding Services for Conversational Question Answering Systems Daniel Braun Adrian Hernandez Mendez Florian Matthes Manfred Langen Technical University of Munich Siemens AG Department of Informatics Corporate Technology daniel.braun,adrian.hernandez,matthes manfred.langen { @tum.de } @siemens.com Abstract Advances in machine learning (ML) • Natural Language Understanding (NLU) as a Conversational interfaces recently gained • a lot of attention. One of the reasons service for the current hype is the fact that chat- In this paper, we focus on the latter. As we bots (one particularly popular form of con- will show in Section2, NLU services are already versational interfaces) nowadays can be used by a number of researchers for building con- created without any programming knowl- versational interfaces. However, due to the lack edge, thanks to different toolkits and so- of a systematic evaluation of theses services, the called Natural Language Understanding decision why one services was prefered over an- (NLU) services. While these NLU ser- other, is usually not well justified. With this paper, vices are already widely used in both, in- we want to bridge this gap and enable both, re- dustry and science, so far, they have not searchers and companies, to make more educated been analysed systematically. In this pa- decisions about which service they should use. We per, we present a method to evaluate the describe the functioning of NLU services and their classification performance of NLU ser- role within the general architecture of chatbots. vices. Moreover, we present two new cor- We explain, how NLU services can be evaluated pora, one consisting of annotated ques- and conduct an evaluation, based on two different tions and one consisting of annotated corpora consisting of nearly 500 annotated ques- questions with the corresponding answers.
    [Show full text]
  • A ROOM with a VUI – VOICE USER INTERFACES in the TESOL CLASSROOM by David Kent Woosong University Daejon, Republic of Korea Dbkent @ Wsu.Ac.Kr
    Teachng Englsh wth Technology, 20(3), 96-124, http://www.tewtjournal.org 96 A ROOM WITH A VUI – VOICE USER INTERFACES IN THE TESOL CLASSROOM by David Kent Woosong University Daejon, Republic of Korea dbkent @ wsu.ac.kr Abstract Disruptive technologies have seen how students interact with their teachers, how we as teachers now prepare and provide learning, and how we might best incorporate artificial intelligence into the classroom. To this end, the pedagogical affordances offered by the voice-user interface of digital assistants is explored. Instructional strategies supported by examples are then provided, along with means for actioning their use in the classroom and evaluating their appropriateness and viability for enhancing language learning. Keywords: digital assistants; voice-user interface; interaction; speaking 1. Introduction As instructors, we no longer talk about technology replacing teachers, or even teachers who use technology replacing those that do not (John & Wheeler, 2015); instead we expect to be incorporating technologies for learning into our classrooms. This means that as instructors today, we need to be able to competently apply that technology and competently assess and evaluate the suitability and appropriateness of how that technology has met intended teaching and learning objectives, while also understanding all levels of the educational potential behind its use, and assisting learners in being able to identify those elements as well (Fotos & Brown, 2004; Levy and Stockwell, 2006). This is important because teaching in the time of digital language learning sees us not just doing old things in new ways, but it has ushered in a total era of ‘newness.’ There are new things to do, new ways to think, new methods of managing relationships with others (and AI – artificial intelligence), and new practices in teaching that require us to adopt new skills and new abilities (Jones & Hafner, 2012).
    [Show full text]
  • Healthcare Chatbot Using Natural Language Processing
    International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 11 | Nov 2020 www.irjet.net p-ISSN: 2395-0072 Healthcare Chatbot using Natural Language Processing Papiya Mahajan1, Rinku Wankhade2, Anup Jawade3, Pragati Dange4, Aishwarya Bhoge5 1,2,3,4,5,Student’s Dept. of Computer Science & Engineering, Dhamangaon Education Society’s College of Engineering and Technology, Dhamangaon Rly. , Maharashtra, India. ---------------------------------------------------------------------***---------------------------------------------------------------------- Abstract - To start a good life healthcare is very important. discovered or vital answer are given or similar answers But it is very difficult to the consult the doctor if any health are displayed can identification which sort of illness you issues. The proposed idea is to create a healthcare chatbot have got supported user symptoms and additionally offers using Natural Language Processing technique it is the part doctor details of explicit illness. It may cut back their of Artificial Intelligence that can diagnose the disease and health problems by victimization this application system. provide basic. To reduce the healthcare costs and improve The system is developed to scale back the tending price accessibility to medical knowledge the Healthcare chatbot is and time of the users because it isn't potential for the built. Some chatbots acts as a medical reference books, users to go to the doctors or consultants once in real time which helps the patient know more about their disease and required. helps to improve their health. The user can achieve the benefit of a healthcare chatbot only when it can diagnose all 2. LITERATURE SURVEY kind of disease and provide necessary information.
    [Show full text]
  • A Comparison of Natural Language Understanding Services to Build a Chatbot in Italian?
    A Comparison of Natural Language Understanding Services to build a chatbot in Italian? Matteo Zubani1;2, Serina Ivan1, and Alfonso Emilio Gerevini1 1 Department of Information Engineering, University of Brescia, Via Branze 38, Brescia 25123, Italy 2 Mega Italia Media S.P.A.,Via Roncadelle 70A, Castel Mella 25030, Italy fm.zubani004,ivan.serina, [email protected] Abstract. All leading IT companies have developed cloud-based plat- forms that allow building a chatbot in few steps and most times without knowledge about programming languages. These services are based on Natural Language Understanding (NLU) engines which deal with identi- fying information such as entities and intents from the sentences provided as input. In order to integrate a chatbot on an e-learning platform, we want to study the performance in intent recognition task of major NLU platforms available on the market through a deep and severe comparison, using an Italian dataset which is provided by the owner of the e-learning platform. We focused on the intent recognition task because we believe that it is the core part of an efficient chatbot, which is able to operate in a complex context with thousands of users who have different language skills. We carried out different experiments and collected performance information about F-score, error rate, response time and robustness of all selected NLU platforms. Keywords: Chatbot · Cloud platform · Natural Language Understand- ing · E-learning. 1 Introduction In the last decade more and more companies have replaced their traditional communication channels with chatbots which can satisfy automatically the users' requests. A chatbot is a virtual person that can talk to a human user using textual messages or voice.
    [Show full text]