CNR-ITD

Conversational applications to support immigrants learning of italian as L2 

External personnel:

Simone Torsani: Language Sciences Researcher, University of Genoa, Department of Modern Languages and Cultures

FRAMEWORK

The problem of the social inclusion of the immigrants is related with the linguistic competencies in the majority language of the arrival countries

The CPIAbot project aims to create and test conversational applications that use mobile devices (normally used by foreigners arriving in our country) and promote the teaching and learning processes of Italian L2 (in and out of classroom ) in a logic of seamless learning, contextualizing the language learning in the places experienced daily by migrants.

OBJECTIVES

• To encourage the use of the Italian language in the classroom as well as outside of them under the assistance of a conversational application.

• To cover a segment of learning not particularly covered by the existing learning applications of Italian L2, specifically that relative to low educated adult immigrants

• To support the teachers during the teaching and learning processes in the classroom, while they explain to all and while they follow their students on an individual basis with exercises of the use of the language.

PLANNED ACTIVITY

• Incremental development of a chatbot calibrated to the needs of Italian L2 learners from PreAlfa level to A2 level.

• Incremental development of a simplified glossary, adapted to the above levels and enriched with multimedia components such as images, animations, videos and audio that can be used by the chatbot to facilitate the language learning process.

• Experimentation in two provincial centers for adult education (CPIA, Genova,Italy), activity that will be followed by about one hundred students.

• Active involvement in the research-action of a dozen CPIA teachers who will support the researchers, both in defining the functional specifications of the chatbot and the use cases, and in its validation in the classroom with the students as the different components of the chatbot and the multimedia glossary will be developed incrementally.

ACTIVITIES

Technological level

Development of a multimodal/multisurface chatbot based on technologies developed ad-hoc in ITD-CNR (among them the NaifJs dialog manager), able to interact vocally and in writing, with examples supported by multimedia content, and with the task of conducting language learning tests in a playful way. Chatbot is:

• Both reactive ("pull" mode), since it responds to the user's stimuli, and proactive ("push" mode), since, by sending messages in natural language, it invites the learner to perform specific actions, even within structured exercises;

• Multimodal, because it allows interaction based on different channels (in input and output) such as voice and written text;

• Multimedial, because the system proposes textual, vocal and visual contents

• Multi-level, because the system outputs are calibrated according to the student's level (set by the teacher);

• Characterized by easy accessibility and learning and by strong student engagement skills inside and outside the classroom.

A simplified chatbot glossary has been created, adapted to the above levels, enriched with multimedia components (images, animations, video and audio) to facilitate the interaction with the conversational application. The multimedia glossary, kit to the chatbot, will be rendered open data, accessible in crowdsourcing mode to encourage its incremental development by anyone who wants to collaborate in its realization and, consequently, its use.

Training plan and assistance to teachers

A support activity for teachers has been carried out to familiarize themselves with the tool and useful for integration in teaching design. An interactive multimedia guide has been created to help teachers to find useful indications for the use of CPIAbot.

The freely accessible and navigable guide offers information and video tutorials on functions and dialogs in the chatbot. 

Following the suspension of teaching activities in all schools of the national territory due to the lockdown COVID-19 (Dpcm 8 mar), activities useful to the use of CPIAbot as technology to support Distance Learning (DaD) have been carried out. Specifically, two actions have been implemented:

• Development of functionalities useful for the creation of virtual classes, the management of messages to students/groups of students and the monitoring of the student interactions

• Assistance to teachers in the design of distance activities supported by CPIAbot (and aimed at illiterate and weakly educated students.)

On the didactic-methodological level

The expected educational results can be traced back to the research questions that the project intends to answer:

1. Significant improvements in the learning process, compared to the use of traditional tools, produced by the interaction with a chatbot through a conversation focused on exploring the terms contained in an integrated multimedia glossary and use cases of everyday life. 

The expected results tied to the significant improvements in the learning process, compared to the use of traditional tools, concern the achievement of the competence objectives defined by the CEFR and the syllabi of reference for the identified levels.

2. The reduced load of the teaching process in the classroom (e.g. in the exhibition and exercise phases as done by the teachers)  produced by acting in a hybrid learning environment where hybridization is at the level of both spaces (classroom, extra-aula, digital) and didactic mediation (teacher in synergy with chatbot). As an indirect result of the project, it is expected a lightening of the workload of teachers during lessons in the classroom, as it is assumed that the tool is:

• An "agile" resource for the retrieval, by the teachers, of simplified contents concerning the vocabulary faced during the lesson (interactive multimedia glossary);

• A support to students' literacy strategies, easily usable by the students themselves (speech transcription and vocal synthesis functions of the written text);

• An automated tool, which allows, within the lesson, personalized interaction to each student during the autonomous exercise phases (exercises accompanying the glossary).

3. Higher level of engagement of the student during the lesson and stimulation to continue the language exercise outside the classroom with activities that the students consider significant and relevant to their life experiences (specifically, the interaction in natural language with an automatic agent that presents real situations) and during which they are stimulated to act with the language in different contexts.

Keywords/Skills

Social inclusion for adult migrants, Adult Literacy, Italian L2 Learning, Mobile Assisted Language Learning, Mobile Learning, Hybrid Learning Spaces, Seamless-learning,  Conversational AI, Conversational agents, Multimodal Chatbots, Voicefirst/Multimodal Virtual Assistants, Mixed-initiative Conversational Agents, Speech-to-text (ASR), Text-to-speech (TTS), Task-oriented Dialog Manager, Node.js, Javascript, Bash/Linux programming. 

Open source/Open data

One of the technological components that enabled the development of CPIABot, the NaifJs dialog manager, was released as open source. The database that constitutes the glossary is released as open data.

SCIENTIFIC PRODUCTION

Torsani S., Robino G., Ravicchio F. (2021). A conversational agent to learn Italian as a Second Language. Invited chapter for the book “CPIAbot: a conversational assistant for learning Italian as second language”, Trentin G. (ed), Nova Science Publishers,  2021.[PDF]

Ravicchio F., Robino R., Torsani S. (2020). Un assistente conversazionale a supporto dell’apprendimento dell’italiano L2 per migranti: CPIAbot. Italian Journal of Educational Technology (IJET). [PDF]

Robino G., Torsani S., Ravicchio F. (2020). Conversational agent in Mobile Assisted Language Learning: CPIAbot. Submitted 05.08.2020 to eLmL 2020 conferenze, 21-25.11.2020, Valencia.[PDF]

Ravicchio F., Robino G., Trentin G, Bernava L. (2019). CPIAbot: un chatbot nell’insegnamento dell’Italiano L2 per stranieri. Atti di Didamatica 2019, Best Paper Award in section: BYOD. Mobile e Mixed Learning (ISBN 978-88-98091-50-8 p. 77-86) [PDF].

PARTICIPATION IN EVENTS / CONFERENCES

Robino, G., Torsani, S., Ravicchio, F. (2020, November 21-25). Conversational agent in Mobile Assisted Language Learning: CPIAbot [Conference session]. eLmL 2020: The Twelfth International Conference on Mobile, Hybrid, and On-line Learning, Valencia, Spain. http://bit.ly/elml20cpiabotpresentation

Trentin, G. Robino, G., Ravicchio, F.(2019, November 15). CPIAbot: a conversational assistant supporting Italian L2 learning for immigrants [Conference session]. C1A0 EXPO – Artificial Intelligence for a Changing Planet. Genoa, Italy. http://bit.ly/itdcnr-c1a0-slides

Ravicchio F., Robino G., Bernava L., Torsani S. (2019, September 21-25). Chatbot e italiano L2. Chatbot e italiano L2 CPIAbot -Un agente automatico per il Pre-A1/A1 [Conference session], Quarto Convegno Internazionale di Linguistica e Glottodidattica Italiana (CILGI). Campobasso, Italy. http://bit.ly/cilgi4cpiabotpresentation

Robino, G. (2018, October 24) Conversational Paradigm. Much more than chatbots [ITD-CNR, internal talk]. Genoa, Italy. www.bit.ly/introforitd

REGISTERED TRADEMARK

The trademark CPIAbot has been registered under number 302020000042412 of June, 3rd, 2020.

COLLABORATIONS

• Digital Humanities phd –University of Genoa 

• Department of Modern Languages and Cultures - University of Genoa

• Genoa Provincial Center for Education of Adults - CPIA Centro-Levante (MIUR)

• Genoa Provincial Center for Education of Adults - CPIA Centro-Ponente (MIUR)

THANK YOU

Since the CPIAbot chatbot was developed as part of a project of the Institute for Educational Technology, in collaboration with the teachers of the Centre-Levante and the Centre-Ponent CPIA, the authors would like to thank the teachers and managers who contributed to it: 

E. Spadaccini, P. Cataldo, G. Rocchi, C. Zappalà, E. Simeoli, A. Gavino, G. Saitto, C. Lazagna, M.Castello, F. Pironi, F. Amato, F. Frosini, G. Grandati, I. Fragale, M. G. Paletta, M. Raggi, M. Ambrosini.

CONTACTS
This email address is being protected from spambots. You need JavaScript enabled to view it.

---
Last update: November 3rd, 2020