Conventional Description of a Human-Machine Interaction on a Collaborative Information Retrieval Task

Jean-Baptiste Louvet (1)

1 MIND - LITIS - Equipe Multi-agent, Interaction, Décision

This document is a translation (from french) the 5 th chpter of the PhD thesis of the author [Lou19] 1. It illustrates the use of the collaborative interaction model (introduced in chapter 3 of the thesis, also introduced in a conference paper [LDC + 17]) created by the author on a collaborative medical information retrieval task. The goal is to show how this model can be applied to a real world problem and how the different tools avaliable adapt to observed needs in concrete examples. We study a human-human interaction corpus to model the problem solving process and then extract the main properties of this process to transpose it as possible behaviors 2 gathered in state tables. Section 1 describes the particularities of collaborative medical information retrieval (CMIR). It also introduces the Cogni-CISMeF corpus, on which our human-human CMIR scenario rely. Following sections develop the stages of this scenario rewritten with the state table formalism: section 2 gives the opening, verbalization and verbalization/terminology alignment stages, section 4 gives the results evaluation stages and section 5 gives the query repair stages. Section 6 describes the modularity levels of our model. Section 7 highlights the propreties of CMIR kept by our model and its limits. Section 8 gives the formal definition of the predicates used in the tables in previous sections.

Domaines: Interface homme-machine [cs.HC]

Jean-Baptiste Louvet. Conventional Description of a Human-Machine Interaction on a Collaborative Information Retrieval Task. [Research Report] INSA Rouen Normandie; LITIS. 2020. ⟨hal-02530158⟩