The influence of persona and conversational task on social interactions with a LLM-controlled embodied conversational agent

Large Language Models (LLMs) have demonstrated remarkable capabilities in conversational tasks. Embodying an LLM as a virtual human allows users to engage in face-to-face social interactions in Virtual Reality. However, the influence of person- and task-related factors in social interactions with LLM-controlled agents remains unclear. In this study, forty-six participants interacted with a virtual agent whose persona was manipulated as extravert or introvert in three different conversational tasks (small talk, knowledge test, convincing). Social-evaluation, emotional experience, and realism were assessed using ratings. Interactive engagement was measured by quantifying participants' words and conversational turns. Finally, we measured participants' willingness to ask the agent for help during the knowledge test. Our findings show that the extraverted agent was more positively evaluated, elicited a more pleasant experience and greater engagement, and was assessed as more realistic compared to the introverted agent. Whereas persona did not affect the tendency to ask for help, participants were generally more confident in the answer when they had help of the LLM. Variation of personality traits of LLM-controlled embodied virtual agents, therefore, affects social-emotional processing and behavior in virtual interactions. Embodied virtual agents allow the presentation of naturalistic social encounters in a virtual environment.

"What can I cook with these ingredients?" -- Understanding cooking-related information needs in conversational search

As conversational search becomes more pervasive, it becomes increasingly important to understand the user's underlying information needs when they converse with such systems in diverse domains. We conduct an in-situ study to understand information needs arising in a home cooking context as well as how they are verbally communicated to an assistant. A human experimenter plays this role in our study. Based on the transcriptions of utterances, we derive a detailed hierarchical taxonomy of diverse information needs occurring in this context, which require different levels of assistance to be solved. The taxonomy shows that needs can be communicated through different linguistic means and require different amounts of context to be understood. In a second contribution we perform classification experiments to determine the feasibility of predicting the type of information need a user has during a dialogue using the turn provided. For this multi-label classification problem, we achieve average F1 measures of 40% using BERT-based models. We demonstrate with examples, which types of need are difficult to predict and show why, concluding that models need to include more context information in order to improve both information need classification and assistance to make such systems usable.

User Preferences and the Shortest Path

Indoor navigation systems leverage shortest path algorithms to calculate routes. In order to define the "shortest path", a cost function has to be specified based on theories and heuristics in the application domain. For the domain of indoor routing, we survey theories and criteria identified in the literature as essential for human path planning. We drive quantitative definitions and integrate them into a cost function that weights each of the criteria separately. We then apply an exhaustive grid search to find weights that lead to an ideal cost function. "Ideal" here is defined as guiding the algorithm to plan routes that are most similar to those chosen by humans. To explore which criteria should be taken into account in an improved pathfinding algorithm, eleven different factors whose favorable impact on route selection has been established in past research were considered. Each factor was included separately in the Dijkstra algorithm and the similarity of thus calculated routes to the actual routes chosen by students at the University of Regensburg was determined. This allows for a quantitative assessment of the factors' impact and further constitutes a way to directly compare them. A reduction of the number of turns, streets, revolving doors, entryways, elevators as well as the combination of the aforementioned factors was found to have a positive effect and generate paths that were favored over the shortest path. Turns and the combination of criteria turned out to be most impactful.

Modelling Users, Intentions, and Structure in Spoken Dialog

We outline how utterances in dialogs can be interpreted using a partial first order logic. We exploit the capability of this logic to talk about the truth status of formulae to define a notion of coherence between utterances and explain how this coherence relation can serve for the construction of AND/OR trees that represent the segmentation of the dialog. In a BDI model we formalize basic assumptions about dialog and cooperative behaviour of participants. These assumptions provide a basis for inferring speech acts from coherence relations between utterances and attitudes of dialog participants. Speech acts prove to be useful for determining dialog segments defined on the notion of completing expectations of dialog participants. Finally, we sketch how explicit segmentation signalled by cue phrases and performatives is covered by our dialog model.

Combining Expression and Content in Domains for Dialog Managers

We present work in progress on abstracting dialog managers from their domain in order to implement a dialog manager development tool which takes (among other data) a domain description as input and delivers a new dialog manager for the described domain as output. Thereby we will focus on two topics; firstly, the construction of domain descriptions with description logics and secondly, the interpretation of utterances in a given domain.

Part-of-Speech-Tagging using morphological information

This paper presents the results of an experiment to decide the question of authenticity of the supposedly spurious Rhesus - a attic tragedy sometimes credited to Euripides. The experiment involves use of statistics in order to test whether significant deviations in the distribution of word categories between Rhesus and the other works of Euripides can or cannot be found. To count frequencies of word categories in the corpus, a part-of-speech-tagger for Greek has been implemented. Some special techniques for reducing the problem of sparse data are used resulting in an accuracy of ca. 96.6%.