La quantité croissante de corpus multimodaux collectés permet de développer de nouvelles méthodes d’analyse de la conversation. Dans la très grande majorité des cas, ces corpus ne comprennent cependant que les enregistrements audio et vidéo, laissant de côté d’autres modalités plus difficiles à récupérer mais apportant un point de vue complémentaire sur la conversation, telle que l’activité cérébrale des locuteurs. Nous présentons donc BrainKT, un corpus de conversation naturelle en français, rassemblant les données audio, vidéo et signaux neurophysiologiques, collecté avec l’objectif d’étudier en profondeur les transmission d’information et l’instanciation du common ground. Pour chacune des conversations des 28 dyades (56 participants), les locuteurs devaient collaborer sur un jeu conversationnel (15min), et étaient ensuite libres de discuter du sujet de leur choix (15min). Pour chaque discussion, les données audio, vidéo, l’activité cérébrale (EEG par Biosemi 64) et physiologique (montre Empatica-E4) sont enregistrées. Cet article situe le corpus dans la littérature, présente le setup expérimental utilisé ainsi les difficultés rencontrées, et les différents niveaux d’annotations proposés pour le corpus.

Les théories de l’interaction suggèrent que l’émergence d’une compréhension mutuelle entre les locuteurs en conversation naturelle dépend de la construction d’une base de connaissances partagée (common ground), mais n’explicitent ni le choix ni les circonstances de la mémorisation de ces informations.Des travaux antérieurs utilisant les métriques dérivées de la théorie de l’information pour analyser la dynamique d’échange d’information ne fournissent pas de moyen efficace de localiser les informations qui entreront dans le common ground. Nous proposons une nouvelle méthode basée sur la segmentation automatique d’une conversation en thèmes qui sont ensuite résumés. L’emplacement des transferts d’informations est finalement obtenu en calculant la distance entre le résumé du thème et les différents énoncés produits par un locuteur. Nous évaluons deux grands modèles de langue (LLMs) sur cette méthode, sur le corpus conversationnel français Paco-Cheese. Plus généralement, nous étudions la façon dont les derniers développement dans le champ des LLMs permettent l’étude de questions s’appuyant normalement fortement sur le jugement d’annotateurs humains.

Interaction theories suggest that the emergence of mutual understanding between speakers in natural conversations depends on the construction of a shared knowledge base (common ground), but the details of which information and the circumstances under which it is memorized are not explained by any model. Previous works have looked at metrics derived from Information Theory to quantify the dynamics of information exchanged between participants, but do not provide an efficient way to locate information that will enter the common ground. We propose a new method based on the segmentation of a conversation into themes followed by their summarization. We then obtain the location of information transfers by computing the distance between the theme summary and the different utterances produced by a speaker. We evaluate two Large Language Models (LLMs) on this pipeline, on the French conversational corpus Paco-Cheese. More generally, we explore how the recent developments in the field of LLMs provide us with the means to implement these new methods and more generally support research into questions that usually heavily relies on human annotators.

An increasing amount of multimodal recordings has been paving the way for the development of a more automatic way to study language and conversational interactions. However this data largely comprises of audio and video recordings, leaving aside other modalities that might complement this external view of the conversation but might be more difficult to collect in naturalistic setups, such as participants brain activity. In this context, we present BrainKT, a natural conversational corpus with audio, video and neuro-physiological signals, collected with the aim of studying information exchanges and common ground instantiation in conversation in a new, more in-depth way. We recorded conversations from 28 dyads (56 participants) during 30 minutes experiments where subjects were first tasked to collaborate on a joint information game, then freely drifted to the topic of their choice. During each session, audio and video were captured, along with the participants’ neural signal (EEG with Biosemi 64) and their electro-physiological activity (with Empatica-E4). The paper situates this new type of resources in the literature, presents the experimental setup and describes the different kinds of annotations considered for the corpus.

The mechanisms underlying human communication have been under investigation for decades, but the answer to how understanding between locutors emerges remains incomplete. Interaction theories suggest the development of a structural alignment between the speakers, allowing for the construction of a shared knowledge base (common ground). In this paper, we propose to apply metrics derived from information theory to quantify the amount of information exchanged between participants, the dynamics of information exchanges, to provide an objective way to measure the common ground instantiation. We focus on a corpus of free conversations augmented with prosodic segmentation and an expert annotation of thematic episodes. We show that during free conversations, the amount of information remains globally constant at the scale of the conversation, but varies depending on the thematic structuring, underlining the role of the speaker introducing the theme. We propose an original methodology applied to uncontrolled material.

We present in this paper the first natural conversation corpus recorded with all modalities and neuro-physiological signals. 5 dyads (10 participants) have been recorded three times, during three sessions (30mns each) with 4 days interval. During each session, audio and video are captured as well as the neural signal (EEG with Emotiv-EPOC) and the electro-physiological one (with Empatica-E4). This resource original in several respects. Technically, it is the first one gathering all these types of data in a natural conversation situation. Moreover, the recording of the same dyads at different periods opens the door to new longitudinal investigations such as the evolution of interlocutors’ alignment during the time. The paper situates this new type of resources with in the literature, presents the experimental setup and describes different annotations enriching the corpus.