Claudio Pinhanez, Paulo Cavalin, Victor Henrique Alves Ribeiro, Ana Appel, Heloisa Candello, Julio Nogima, Mauro Pichiliani, Melina Guerra, Maira de Bayser, Gabriel Malfatti, Henrique Ferreira. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

This paper describes how machine learning training data and symbolic knowledge from curators of conversational systems can be used together to improve the accuracy of those systems and to enable better curatorial tools. This is done in the context of a real-world practice of curators of conversational systems who often embed taxonomically-structured meta-knowledge into their documentation. The paper provides evidence that the practice is quite common among curators, that is used as part of their collaborative practices, and that the embedded knowledge can be mined by algorithms. Further, this meta-knowledge can be integrated, using neuro-symbolic algorithms, to the machine learning-based conversational system, to improve its run-time accuracy and to enable tools to support curatorial tasks. Those results point towards new ways of designing development tools which explore an integrated use of code and documentation by machines.

How does the presence of an audience influence the social interaction with a conversational system in a physical space? To answer this question, we analyzed data from an art exhibit where visitors interacted in natural language with three chatbots representing characters from a book. We performed two studies to explore the influence of audiences. In Study 1, we did fieldwork cross-analyzing the reported perception of the social interaction, the audience conditions (visitor is alone, visitor is observed by acquaintances and/or strangers), and control variables such as the visitor's familiarity with the book and gender. In Study 2, we analyzed over 5,000 conversation logs and video recordings, identifying dialogue patterns and how they correlated with the audience conditions. Some significant effects were found, suggesting that conversational systems in physical spaces should be designed based on whether other people observe the user or not.

In an offering of CS 1, we monitored the amount of help students in a class received through various means. We found that most students wanted help that went beyond office hours and email, and that for the vast majority of them, their grades correlated positively with the amount of help they received. However, increasing help also requires increasing instructional resources unless techniques can be found to separate the surmountable difficulties from the insurmountable ones. We have developed such a technique that looks at individual interaction logs and classifies difficulties incrementally as they occur. The insight behind the technique is that when students face surmountable difficulties, they tend to repeat certain action sequences, which can be detected to distinguish the surmountable difficulties from the insurmountable ones. A lab study shows that the mechanism gives significantly better results than the baselines.

Mobile applications, which include calendars, browsers, and text editors, are part of our lives nowadays. Most of the mobile applications are single user, i.e. they do not allow the cooperative work of group of users simultaneously. Adaptation is a technique to transform single user applica

Tabletop can be used in groupware systems in order to both provide a rich interface shared workspaces and support work with co-located and remote users. Tabletop-based groupware systems can allow users to work together over a shared horizontal display. Music composition is an application that can ex

Collaborative editing systems often use communication channels in order to make the collaboration more effective. In this work, we present a case study of pair communication using audio via VoIP (Voice over Internet Protocol) technology during collaborative modeling sessions. The study provides an analysis of audio and textual chat as communication medium and presents data on usage patterns, user interaction and attitudes when using collaborative editing systems. The qualitative and quantitative analysis suggests that VoIP technology does have advantages over textual chat in collaborative modeling when used for communication.

Collaborative editing systems require mechanisms of control in order to make the collaboration more effective. The synchronous collaboration allows the notion of simultaneity; however, it demands more adequate perception elements. In the collaborative edition of UML diagrams where dependencies between their modeling elements do exist, it is necessary to apply a mechanism of control to support the edition of consistent models. In this work, we present a multi-level locking mechanism that automatically manages locks in the modeling elements of UML diagrams during a collaborative session. This paper also describes a controlled experiment conducted to study the use of the multi-level locking mechanism during collaborative modeling sessions, and the observations from the empirical data collected.

Building a collaborative application from scratch is a hard task. In the last decade many advances have been made to help the developers to construct collaborative applications, however little effort has been made to extend existing single-user applications to support real-time collaboration. This work presents a mapping from the main components of an existing single-user model-view-controller based application to multiuser real-time components of the collaborative application. The mapping allows reuse of existing single-user components by facilitating the construction of collaborative applications. This paper describes the mapping, the extension of an existing single-user application and discusses an experiment of the prototype application where the mapping was applied