TY - CPAPER TI - Conv-BDI: An Extension of the BDI Framework for Conversational Agents AU - Walker, Nicholas Thomas AU - Wagner, Nicolas AU - Hilgendorf, Laetitia AU - Ultes, Stefan PY - 2026 SP - 104–114 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/114436 LA - eng AB - With the large and increasing variety of architectures for conversational agent design, there is a need to investigate the necessary elements of practical conversational agents in light of new technologies. To address this need, we introduce a new abstract framework of conversational agents design which we call Conv-BDI. The BDI model is a longestablished theory of decision-making in artificial agents, which the Conv-BDI model extends to describe the design of conversational agents from traditional symbolic logic-based models or statistical models to more recent LLM-based agents. Specifically, we extend the core BDI model with notions of Purpose and Behavioral Guidelines, while also elaborating on the role of system actions within this framework. The Conv-BDI model thus provides a framework of intentionality in conversational agents that can be applied to to design of contemporary conversational agents. KW - - ER - TY - CPAPER TI - Enhancing Model Transparency : A Dialogue System Approach to XAI with Domain Knowledge AU - Feustel, Isabel AU - Rach, Niklas AU - Minker, Wolfgang AU - Ultes, Stefan PY - 2026 SP - 248-258 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/114438 LA - eng AB - Explainable artifcial intelligence (XAI) is a rapidly evolving feld that seeks to create AI systems that can provide humanunderstandable explanations for their decisionmaking processes. However, these explanations rely on model and data-specifc information only. To support better human decisionmaking, integrating domain knowledge into AI systems is expected to enhance understanding and transparency. In this paper, we present an approach for combining XAI explanations with domain knowledge within a dialogue system. We concentrate on techniques derived from the feld of computational argumentation to incorporate domain knowledge and corresponding explanations into human-machine dialogue. We implement the approach in a prototype system for an initial user evaluation, where users interacted with the dialogue system to receive predictions from an underlying AI model. The participants were able to explore different types of explanations and domain knowledge. Our results indicate that users tend to more effectively evaluate model performance when domain knowledge is integrated. On the other hand, we found that domain knowledge was not frequently requested by the user during dialogue interactions. KW - - ER - TY - CPAPER TI - On the Controllability of Large Language Models for Dialogue Interaction AU - Wagner, Nicolas AU - Ultes, Stefan PY - 2026 SP - 216-221 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/114437 LA - eng AB - This paper investigates the enhancement of Dialogue Systems by integrating the creative capabilities of Large Language Models. While traditional Dialogue Systems focus on understanding user input and selecting appropriate system actions, Language Models excel at generating natural language text based on prompts. Therefore, we propose to improve controllability and coherence of interactions by guiding a Language Model with control signals that enable explicit control over the system behaviour. To address this, we tested and evaluated our concept in 815 conversations with over 3600 dialogue exchanges on a dataset. Our experiment examined the quality of generated system responses using two strategies: An unguided strategy where task data was provided to the models, and a controlled strategy in which a simulated Dialogue Controller provided appropriate system actions. The results show that the average BLEU score and the classifcation of dialogue acts improved in the controlled Natural Language Generation. KW - - ER - TY - CPAPER TI - The Influence of Avatar Interfaces on Argumentative Dialogues AU - Aicher, Annalena AU - Weber, Klaus AU - André, Elisabeth AU - Minker, Wolfgang AU - Ultes, Stefan PY - 2025 SP - 1-8 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/110658 LA - eng AB - Humans form opinions and justify different points of view by ex- changing arguments and knowledge. Likewise to human-human interaction, the way arguments are presented infuence the user’s willingness to engage into a critical refection. Especially when interacting with conversational agents the user’s engagement and motivation are important factors and highly infuence the success or failure of such a mixed team. To maintain the users’ trust and satis- faction, the users’ perception of the respective system is an important indicator. Thus, this work investigates the design of a cooperative argumentative dialogue system using a virtual avatar compared to a non-avatar interface by evaluating a crowdsourcing study conducted with 84 participants. The results indicate, that the avatar system is perceived as signifcantly more appealing and natural and thus, engaging which also infuences the acceptance and perception of the quality of presented arguments. Furthermore, we found that the presence of the avatar often led to an increase in the anticipated level of conversational profciency similar to that of a human interlocutor. Therefore, this work provides important insights for the design of future cooperative argumentative virtual avatar interfaces. KW - Conversational Engagement KW - User Trust KW - Avatar Interface KW - Human-Computer Interaction KW - Crowdsourcing Study ER - TY - CPAPER TI - BEA : Building Engaging Argumentation AU - Aicher, Annalena AU - Weber, Klaus AU - André, Elisabeth AU - Minker, Wolfgang AU - Ultes, Stefan PY - 2025 SP - 279-295 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/110816 LA - eng AB - Exchanging arguments and knowledge in conversations is an intuitive way for humans to form opinions and reconcile opposing viewpoints. The vast amount of information available on the internet, often accessed through search engines, presents a considerable challenge. Managing and filtering this overwhelming wealth of data raises the potential for intellectual isolation. This can stem either from personalized searches that create “filter bubbles” by considering a user’s history and preferences, or from the intrinsic, albeit unconscious, tendency of users to seek information that aligns with their existing beliefs, forming “self-imposed filter bubbles”. To address this issue, we introduce a model aimed at engaging the user in a critical examination of presented arguments and propose the use of a virtual agent engaging in a deliberative dialogue with human users to facilitate a fair and unbiased opinion formation. Our experiments have demonstrated the success of these models and their implementation. As a result, this work offers valuable insights for the design of future cooperative argumentative dialogue systems. KW - Cooperative Argumentative Dialogue Systems KW - Reflective (User) Engagement (RUE) KW - Conversational Engagement KW - User Attention KW - User Focus KW - Gaze Tracking KW - Virtual Avatar ER - TY - CPAPER TI - Conv-BDI: An Extension of the BDI Framework for Conversational Agents AU - Walker, Nicholas Thomas AU - Wagner, Nicolas AU - Hilgendorf, Laetitia AU - Ultes, Stefan PY - 2025 T2 - Proceedings of the 29th Workshop on the Semantics and Pragmatics of Dialogue – Full Papers SP - 104–114 PB - SEMDIAL CY - Bielefeld UR - https://fis.uni-bamberg.de/handle/uniba/110385 LA - eng AB - With the large and increasing variety of architectures for conversational agent design, there is a need to investigate the necessary elements of practical conversational agents in light of new technologies. To address this need, we introduce a new abstract framework of conversational agents design which we call Conv-BDI. The BDI model is a longestablished theory of decision-making in artificial agents, which the Conv-BDI model extends to describe the design of conversational agents from traditional symbolic logic-based models or statistical models to more recent LLM-based agents. Specifically, we extend the core BDI model with notions of Purpose and Behavioral Guidelines, while also elaborating on the role of system actions within this framework. The Conv-BDI model thus provides a framework of intentionality in conversational agents that can be applied to to design of contemporary conversational agents. KW - - ER - TY - JOUR TI - Enhancing Reflective and Conversational User Engagement in Argumentative Dialogues with Virtual Agents AU - Aicher, Annalena AU - Matsuda, Yuki AU - Yasumoto, Keichii AU - Minker, Wolfgang AU - André, Elisabeth AU - Ultes, Stefan PY - 2025 SP - 1-44 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/107706 LA - eng AB - In their process of information seeking, human users tend to selectively ignore information that contradicts their pre-existing beliefs or opinions. These so-called “self-imposed flter bubbles” (SFBs) pose a signifcant challenge for argumentative conversational agents aiming to facilitate critical, unbiased opinion formation on controversial topics. With the ultimate goal of developing a system that helps users break their self-imposed flter bubbles (SFBs), this paper aims to investigate the role of co-speech gestures, specifcally examining how these gestures signifcantly contribute to achieving this objective. This paper extends current research by examining methods to engage users in cooperative discussions with a virtual human-like agent, encouraging a deep refection on arguments to disrupt SFBs. Specifcally, we investigate the agent’s non-verbal behavior in the form of co-speech gestures. We analyze whether co-speech gestures, depending on the conveyed information, enhance motivation, and thus conversational user engagement, thereby encouraging users to consider information that could potentially disrupt their SFBs. The fndings of a laboratory study with 56 participants highlight the importance of non-verbal agent behaviors, such as co-speech gestures, in improving users’ perceptions of the interaction and the conveyed content. This effect is particularly notable when the content aims to challenge the user’s SFB. Therefore, this research offers valuable insights into enhancing user engagement in the design of multimodal interactions with future cooperative argumentative virtual agents. KW - non-verbal agent behavior KW - virtual agents KW - co-speech gestures KW - user interest KW - confirmation bias KW - self-imposed filter bubbles (SFB) KW - argumentative dialogue systems (ADS) KW - human–agent interaction ER - TY - CPAPER TI - Exploring the Impact of Non-Verbal Virtual Agent Behavior on User Engagement in Argumentative Dialogues AU - Aicher, Annalena Bea AU - Matsuda, Yuki AU - Yasumoto, Keichii AU - Minker, Wolfgang AU - André, Elisabeth AU - Ultes, Stefan PY - 2024 PB - arXiv DO - 10.48550/arxiv.2411.11102 UR - https://fis.uni-bamberg.de/handle/uniba/110388 LA - eng ER - TY - CPAPER TI - Enhancing Model Transparency : A Dialogue System Approach to XAI with Domain Knowledge AU - Feustel, Isabel AU - Rach, Niklas AU - Minker, Wolfgang AU - Ultes, Stefan PY - 2024 T2 - Proceedings of the 25th Annual Meeting of the Special Interest Group on Discourse and Dialogue SP - 248-258 PB - Association for Computational Linguistics CY - Kyoto DO - 10.18653/v1/2024.sigdial-1.22 UR - https://fis.uni-bamberg.de/handle/uniba/110387 LA - eng AB - Explainable artifcial intelligence (XAI) is a rapidly evolving feld that seeks to create AI systems that can provide humanunderstandable explanations for their decisionmaking processes. However, these explanations rely on model and data-specifc information only. To support better human decisionmaking, integrating domain knowledge into AI systems is expected to enhance understanding and transparency. In this paper, we present an approach for combining XAI explanations with domain knowledge within a dialogue system. We concentrate on techniques derived from the feld of computational argumentation to incorporate domain knowledge and corresponding explanations into human-machine dialogue. We implement the approach in a prototype system for an initial user evaluation, where users interacted with the dialogue system to receive predictions from an underlying AI model. The participants were able to explore different types of explanations and domain knowledge. Our results indicate that users tend to more effectively evaluate model performance when domain knowledge is integrated. On the other hand, we found that domain knowledge was not frequently requested by the user during dialogue interactions. KW - - ER - TY - CPAPER TI - On the Controllability of Large Language Models for Dialogue Interaction AU - Wagner, Nicolas AU - Ultes, Stefan PY - 2024 T2 - Proceedings of the 25th Annual Meeting of the Special Interest Group on Discourse and Dialogue SP - 216-221 PB - Association for Computational Linguistics CY - Kyoto DO - 10.18653/v1/2024.sigdial-1.19 UR - https://fis.uni-bamberg.de/handle/uniba/110386 LA - eng AB - This paper investigates the enhancement of Dialogue Systems by integrating the creative capabilities of Large Language Models. While traditional Dialogue Systems focus on understanding user input and selecting appropriate system actions, Language Models excel at generating natural language text based on prompts. Therefore, we propose to improve controllability and coherence of interactions by guiding a Language Model with control signals that enable explicit control over the system behaviour. To address this, we tested and evaluated our concept in 815 conversations with over 3600 dialogue exchanges on a dataset. Our experiment examined the quality of generated system responses using two strategies: An unguided strategy where task data was provided to the models, and a controlled strategy in which a simulated Dialogue Controller provided appropriate system actions. The results show that the average BLEU score and the classifcation of dialogue acts improved in the controlled Natural Language Generation. KW - - ER - TY - CPAPER TI - Proceedings of the 25th Annual Meeting of the Special Interest Group on Discourse and Dialogue ED - Kawahara, Tatsuya ED - Demberg, Vera ED - Ultes, Stefan ED - Inoue, Koji ED - Mehri, Shikib ED - Howcroft, David ED - Komatani, Kazunori PY - 2024 PB - Association for Computational Linguistics CY - Kyoto, Japan DO - 10.18653/v1/2024.sigdial-1 UR - https://fis.uni-bamberg.de/handle/uniba/110384 LA - eng ER - TY - JOUR TI - Enhancing Reflective and Conversational User Engagement in Argumentative Dialogues with Virtual Agents AU - Aicher, Annalena AU - Matsuda, Yuki AU - Yasumoto, Keichii AU - Minker, Wolfgang AU - André, Elisabeth AU - Ultes, Stefan PY - 2024 JO - Multimodal technologies and interaction VL - 8 IS - 71 PB - MDPI AG CY - Basel SN - 2414-4088 DO - 10.3390/mti8080071 UR - https://fis.uni-bamberg.de/handle/uniba/97134 LA - eng AB - In their process of information seeking, human users tend to selectively ignore information that contradicts their pre-existing beliefs or opinions. These so-called “self-imposed flter bubbles” (SFBs) pose a signifcant challenge for argumentative conversational agents aiming to facilitate critical, unbiased opinion formation on controversial topics. With the ultimate goal of developing a system that helps users break their self-imposed flter bubbles (SFBs), this paper aims to investigate the role of co-speech gestures, specifcally examining how these gestures signifcantly contribute to achieving this objective. This paper extends current research by examining methods to engage users in cooperative discussions with a virtual human-like agent, encouraging a deep refection on arguments to disrupt SFBs. Specifcally, we investigate the agent’s non-verbal behavior in the form of co-speech gestures. We analyze whether co-speech gestures, depending on the conveyed information, enhance motivation, and thus conversational user engagement, thereby encouraging users to consider information that could potentially disrupt their SFBs. The fndings of a laboratory study with 56 participants highlight the importance of non-verbal agent behaviors, such as co-speech gestures, in improving users’ perceptions of the interaction and the conveyed content. This effect is particularly notable when the content aims to challenge the user’s SFB. Therefore, this research offers valuable insights into enhancing user engagement in the design of multimodal interactions with future cooperative argumentative virtual agents. KW - non-verbal agent behavior KW - virtual agents KW - co-speech gestures KW - user interest KW - confirmation bias KW - self-imposed filter bubbles (SFB) KW - argumentative dialogue systems (ADS) KW - human–agent interaction ER - TY - CPAPER TI - BEA : Building Engaging Argumentation AU - Aicher, Annalena AU - Weber, Klaus AU - André, Elisabeth AU - Minker, Wolfgang AU - Ultes, Stefan PY - 2024 T2 - Robust Argumentation Machines : First International Conference, RATIO 2024, Bielefeld, Germany, June 5–7, 2024, Proceedings SP - 279-295 PB - Springer Nature Switzerland CY - Cham SN - 978-3-031-63535-9 DO - 10.1007/978-3-031-63536-6_17 UR - https://fis.uni-bamberg.de/handle/uniba/96623 LA - eng AB - Exchanging arguments and knowledge in conversations is an intuitive way for humans to form opinions and reconcile opposing viewpoints. The vast amount of information available on the internet, often accessed through search engines, presents a considerable challenge. Managing and filtering this overwhelming wealth of data raises the potential for intellectual isolation. This can stem either from personalized searches that create “filter bubbles” by considering a user’s history and preferences, or from the intrinsic, albeit unconscious, tendency of users to seek information that aligns with their existing beliefs, forming “self-imposed filter bubbles”. To address this issue, we introduce a model aimed at engaging the user in a critical examination of presented arguments and propose the use of a virtual agent engaging in a deliberative dialogue with human users to facilitate a fair and unbiased opinion formation. Our experiments have demonstrated the success of these models and their implementation. As a result, this work offers valuable insights for the design of future cooperative argumentative dialogue systems. KW - Cooperative Argumentative Dialogue Systems KW - Reflective (User) Engagement (RUE) KW - Conversational Engagement KW - User Attention KW - User Focus KW - Gaze Tracking KW - Virtual Avatar ER - TY - CPAPER TI - Towards Breaking the Self-imposed Filter Bubble in Argumentative Dialogues AU - Aicher, Annalena AU - Kornmueller, Daniel AU - Matsuda, Yuki AU - Ultes, Stefan AU - Minker, Wolfgang AU - Yasumoto, Keiichi PY - 2024 SP - 593–604 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/94787 LA - eng AB - Human users tend to selectively ignore information that contradicts their pre-existing beliefs or opinions in their process of information seeking. These “self-imposed filter bubbles” (SFB) pose a significant challenge for cooperative argumentative dialogue systems aiming to build an unbiased opinion and a better understanding of the topic at hand. To address this issue, we develop a strategy for overcoming users’ SFB within the course of the interaction. By continuously modeling the user’s position in relation to the SFB, we are able to identify the respective arguments which maximize the probability to get outside the SFB and present them to the user. We implemented this approach in an argumentative dialogue system and evaluated in a laboratory user study with 60 participants to show its validity and applicability. The findings suggest that the strategy was successful in breaking users’ SFBs and promoting a more reflective and comprehensive discussion of the topic. KW - Self-imposed Filter Bubble ER - TY - CPAPER TI - Self-imposed Filter Bubble Model for Argumentative Dialogues AU - Aicher, Annalena AU - Kornmüller, Daniel AU - Minker, Wolfgang AU - Ultes, Stefan PY - 2024 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/93277 LA - eng AB - During their information seeking people tend to filter out all the parts of the available information that do not fit their existing beliefs or opinions. In this paper we present a model for this “Self-imposed Filter Bubble” (SFB) consisting of four dimensions. Thereby, we aim to 1) estimate the probability of the user being caught in an SFB and consequently, 2) identify suitable clues to reduce this probability in the further course of a dialogue. Using an exemplary implementation in an argumentative dialogue system, we demonstrate the validity and applicability of this model in an online user study with 102 participants. These findings serve as a basis for developing a system strategy to break the user’s SFB and contribute to a sustainable and profound reflection on a topic from all viewpoints. KW - Confirmation Bias KW - Echo Chambers KW - User Modeling KW - Computational Argumentation KW - Cooperative Argumentative Dialogue Systems (ADS) ER - TY - CPAPER TI - The Influence of Avatar Interfaces on Argumentative Dialogues AU - Aicher, Annalena AU - Weber, Klaus AU - André, Elisabeth AU - Minker, Wolfgang AU - Ultes, Stefan PY - 2023 T2 - IVA '23: Proceedings of the 23rd ACM International Conference on Intelligent Virtual Agents SP - 1-8 PB - Association for Computing Machinery CY - New York SN - 978-1-4503-9994-4 DO - 10.1145/3570945.3607343 UR - https://fis.uni-bamberg.de/handle/uniba/96550 LA - eng AB - Humans form opinions and justify different points of view by ex- changing arguments and knowledge. Likewise to human-human interaction, the way arguments are presented infuence the user’s willingness to engage into a critical refection. Especially when interacting with conversational agents the user’s engagement and motivation are important factors and highly infuence the success or failure of such a mixed team. To maintain the users’ trust and satis- faction, the users’ perception of the respective system is an important indicator. Thus, this work investigates the design of a cooperative argumentative dialogue system using a virtual avatar compared to a non-avatar interface by evaluating a crowdsourcing study conducted with 84 participants. The results indicate, that the avatar system is perceived as signifcantly more appealing and natural and thus, engaging which also infuences the acceptance and perception of the quality of presented arguments. Furthermore, we found that the presence of the avatar often led to an increase in the anticipated level of conversational profciency similar to that of a human interlocutor. Therefore, this work provides important insights for the design of future cooperative argumentative virtual avatar interfaces. KW - Conversational Engagement KW - User Trust KW - Avatar Interface KW - Human-Computer Interaction KW - Crowdsourcing Study ER - TY - CPAPER TI - Evaluating a Spoken Argumentative Dialogue System AU - Aicher, Annalena AU - Hillmann, Stefan AU - Michael, Thilo AU - Möller, Sebastian AU - Minker, Wolfgang AU - Ultes, Stefan PY - 2023 T2 - Social Computing and Social Media Social Computing and Social Media : 15th International Conference, SCSM 2023, Held as Part of the 25th HCI International Conference, HCII 2023, Copenhagen, Denmark, July 23–28, 2023, Proceedings SP - 425–434 PB - Springer Nature Switzerland CY - Cham SN - 978-3-031-35927-9 DO - 10.1007/978-3-031-35927-9_29 UR - https://fis.uni-bamberg.de/handle/uniba/91882 LA - eng ER - TY - CPAPER TI - Towards Breaking the Self-imposed Filter Bubble in Argumentative Dialogues AU - Aicher, Annalena AU - Kornmueller, Daniel AU - Matsuda, Yuki AU - Ultes, Stefan AU - Minker, Wolfgang AU - Yasumoto, Keiichi PY - 2023 T2 - Proceedings of the 24th Meeting of the Special Interest Group on Discourse and Dialogue SP - 593–604 PB - Association for Computational Linguistics CY - Prague, Czechia DO - 10.18653/v1/2023.sigdial-1.56 UR - https://fis.uni-bamberg.de/handle/uniba/91881 LA - eng AB - Human users tend to selectively ignore information that contradicts their pre-existing beliefs or opinions in their process of information seeking. These “self-imposed filter bubbles” (SFB) pose a significant challenge for cooperative argumentative dialogue systems aiming to build an unbiased opinion and a better understanding of the topic at hand. To address this issue, we develop a strategy for overcoming users’ SFB within the course of the interaction. By continuously modeling the user’s position in relation to the SFB, we are able to identify the respective arguments which maximize the probability to get outside the SFB and present them to the user. We implemented this approach in an argumentative dialogue system and evaluated in a laboratory user study with 60 participants to show its validity and applicability. The findings suggest that the strategy was successful in breaking users’ SFBs and promoting a more reflective and comprehensive discussion of the topic. KW - Self-imposed Filter Bubble ER - TY - CPAPER TI - Towards Modelling Elaborateness in Argumentative Dialogue Systems AU - Aicher, Annalena AU - Fuchs, Marc AU - Minker, Wolfgang AU - Ultes, Stefan ED - Degen, Helmut ED - Ntoa, Stavroula PY - 2023 T2 - Artificial Intelligence in HCI : 4th International Conference, AI-HCI 2023, Held as Part of the 25th HCI International Conference, HCII 2023, Copenhagen, Denmark, July 23–28, 2023, Proceedings, Part II SP - 3–22 PB - Springer Nature Switzerland CY - Cham SN - 978-3-031-35894-4 DO - 10.1007/978-3-031-35894-4_1 UR - https://fis.uni-bamberg.de/handle/uniba/91880 LA - eng ER - TY - CPAPER TI - Unified Conversational Models with System-Initiated Transitions between Chit-Chat and Task-Oriented Dialogues AU - Liu, Ye AU - Ultes, Stefan AU - Minker, Wolfgang AU - Maier, Wolfgang PY - 2023 T2 - CUI '23: Proceedings of the 5th International Conference on Conversational User Interfaces SP - 1-9 PB - Association for Computing Machinery CY - New York NY, United States SN - 979-8-4007-0014-9 DO - 10.1145/3571884.3597125 UR - https://fis.uni-bamberg.de/handle/uniba/91841 LA - eng ER - TY - CPAPER TI - Self-imposed Filter Bubble Model for Argumentative Dialogues AU - Aicher, Annalena AU - Kornmüller, Daniel AU - Minker, Wolfgang AU - Ultes, Stefan PY - 2023 T2 - Proceedings of the 5th International Conference on Conversational User Interfaces PB - ACM CY - New York,NY, United States SN - 979-8-4007-0014-9 DO - 10.1145/3571884.3597131 UR - https://fis.uni-bamberg.de/handle/uniba/91843 LA - eng AB - During their information seeking people tend to filter out all the parts of the available information that do not fit their existing beliefs or opinions. In this paper we present a model for this “Self-imposed Filter Bubble” (SFB) consisting of four dimensions. Thereby, we aim to 1) estimate the probability of the user being caught in an SFB and consequently, 2) identify suitable clues to reduce this probability in the further course of a dialogue. Using an exemplary implementation in an argumentative dialogue system, we demonstrate the validity and applicability of this model in an online user study with 102 participants. These findings serve as a basis for developing a system strategy to break the user’s SFB and contribute to a sustainable and profound reflection on a topic from all viewpoints. KW - Confirmation Bias KW - Echo Chambers KW - User Modeling KW - Computational Argumentation KW - Cooperative Argumentative Dialogue Systems (ADS) ER - TY - CPAPER TI - Exploring the Mobile Usability of Argumentative Dialogue Systems for Opinion Building AU - Aicher, Annalena AU - Hillmann, Stefan AU - Möller, Sebastian AU - Minker, Wolfgang AU - Ultes, Stefan ED - Salvendy, Gavriel PY - 2023 T2 - Design, operation and evaluation of mobile communications : 4th international conference, MOBILE 2023, held as part of the 25th HCI International Conference, HCII 2023, Copenhagen, Denmark, July 23–28, 2023, proceedings SP - 127-143 PB - Springer CY - Cham, Switzerland SN - 978-3-031-35920-0 DO - 10.1007/978-3-031-35921-7_9 UR - https://fis.uni-bamberg.de/handle/uniba/91844 LA - eng ER - TY - CPAPER TI - System-Initiated Transitions from Chit-Chat to Task-Oriented Dialogues with Transition Info Extractor and Transition Sentence Generator AU - Liu, Ye AU - Ultes, Stefan AU - Minker, Wolfgang AU - Maier, Wolfgang PY - 2023 SP - 279–292 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/91575 LA - eng AB - In this work, we study dialogue scenarios that start from chit-chat but eventually switch to task-related services, and investigate how a unified dialogue model, which can engage in both chit-chat and task-oriented dialogues, takes the initiative during the dialogue mode transition from chit-chat to task-oriented in a coherent and cooperative manner. We firstly build a transition info extractor (TIE) that keeps track of the preceding chit-chat interaction and detects the potential user intention to switch to a task-oriented service. Meanwhile, in the unified model, a transition sentence generator (TSG) is extended through efficient Adapter tuning and transition prompt learning. When the TIE successfully finds task-related information from the preceding chit-chat, such as a transition domain (“train” in Figure fig: system-initiated transition from chit-chat to task-oriented.), then the TSG is activated automatically in the unified model to initiate this transition by generating a transition sentence under the guidance of transition information extracted by TIE. The experimental results show promising performance regarding the proactive transitions. We achieve an additional large improvement on TIE model by utilizing Conditional Random Fields (CRF). The TSG can flexibly generate transition sentences while maintaining the unified capabilities of normal chit-chat and task-oriented response generation. KW - System-Initiated Transitions ER - TY - CPAPER TI - Unified Conversational Models with System-Initiated Transitions between Chit-Chat and Task-Oriented Dialogues AU - Liu, Ye AU - Minker, Wolfgang AU - Maier, Wolfgang AU - Ultes, Stefan PY - 2023 T2 - CUI '23 : Proceedings of the 5th International Conference on Conversational User Interfaces SP - 1-9 PB - Association for Computing Machinery CY - New York, NY, USA SN - 9798400700149 DO - 10.1145/3571884.3597125 UR - https://fis.uni-bamberg.de/handle/uniba/91204 LA - eng ER - TY - CPAPER TI - System-Initiated Transitions from Chit-Chat to Task-Oriented Dialogues with Transition Info Extractor and Transition Sentence Generator AU - Liu, Ye AU - Ultes, Stefan AU - Minker, Wolfgang AU - Maier, Wolfgang PY - 2023 T2 - Proceedings of the 16th International Natural Language Generation Conference SP - 279–292 PB - Association for Computational Linguistics CY - Prague, Czechia UR - https://fis.uni-bamberg.de/handle/uniba/91203 LA - eng AB - In this work, we study dialogue scenarios that start from chit-chat but eventually switch to task-related services, and investigate how a unified dialogue model, which can engage in both chit-chat and task-oriented dialogues, takes the initiative during the dialogue mode transition from chit-chat to task-oriented in a coherent and cooperative manner. We firstly build a transition info extractor (TIE) that keeps track of the preceding chit-chat interaction and detects the potential user intention to switch to a task-oriented service. Meanwhile, in the unified model, a transition sentence generator (TSG) is extended through efficient Adapter tuning and transition prompt learning. When the TIE successfully finds task-related information from the preceding chit-chat, such as a transition domain (“train” in Figure fig: system-initiated transition from chit-chat to task-oriented.), then the TSG is activated automatically in the unified model to initiate this transition by generating a transition sentence under the guidance of transition information extracted by TIE. The experimental results show promising performance regarding the proactive transitions. We achieve an additional large improvement on TIE model by utilizing Conditional Random Fields (CRF). The TSG can flexibly generate transition sentences while maintaining the unified capabilities of normal chit-chat and task-oriented response generation. KW - System-Initiated Transitions ER - TY - CPAPER TI - GraphWOZ: Dialogue Management with Conversational Knowledge Graphs AU - Walker, Nicholas Thomas AU - Ultes, Stefan AU - Lison, Pierre PY - 2023 SP - 1-17 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/59668 LA - eng AB - We present a new approach to dialogue management using conversational knowledge graphs as core representation of the dialogue state. To this end, we introduce a new dataset, GraphWOZ, which comprises Wizard-of-Oz dialogues in which human participants interact with a robot acting as a receptionist. In contrast to most existing work on dialogue management, GraphWOZ relies on a dialogue state explicitly represented as a dynamic knowledge graph instead of a fixed set of slots. This graph is composed of a varying number of entities (such as individuals, places, events, utterances and mentions) and relations between them (such as persons being part of a group or attending an event). The graph is then regularly updated on the basis of new observations and system actions. GraphWOZ is released along with detailed manual annotations related to the user intents, system responses, and reference relations occurring in both user and system turns. Based on GraphWOZ, we present experimental results for two dialogue management tasks, namely conversational entity linking and response ranking. For conversational entity linking, we show how to connect utterance mentions to their corresponding entity in the knowledge graph with a neural model relying on a combination of both string and graph-based features. Response ranking is then performed by summarizing the relevant content of the graph into a text, which is concatenated with the dialogue history and employed as input to score possible responses to a given dialogue state. KW - Dialogue Management KW - Knowledge Graph ER - TY - JOUR TI - Natural language understanding for argumentative dialogue systems in the opinion building domain AU - Abro, Waheed Ahmed AU - Aicher, Annalena AU - Rach, Niklas AU - Ultes, Stefan AU - Minker, Wolfgang AU - Qi, Guilin PY - 2022 JO - Knowledge-Based Systems VL - 242 SN - 0950-7051 DO - 10.1016/j.knosys.2022.108318 UR - https://fis.uni-bamberg.de/handle/uniba/59486 LA - eng ER - TY - CHAP TI - Empathetic Dialogue Generation with Pre-trained RoBERTa-GPT2 and External Knowledge AU - Liu, Ye AU - Maier, Wolfgang AU - Ultes, Stefan AU - Minker, Wolfgang PY - 2022 T2 - Conversational AI for Natural Human-Centric Interaction SP - 67-81 PB - Springer Nature CY - Singapore DO - 10.1007/978-981-19-5538-9_5 UR - https://fis.uni-bamberg.de/handle/uniba/59484 LA - eng ER - TY - JOUR TI - When to Say What and How: Adapting the Elaborateness and Indirectness of Spoken Dialogue Systems AU - Miehle, Juliana AU - Minker, Wolfgang AU - Ultes, Stefan PY - 2022 JO - Dialogue & Discourse VL - 13 IS - 1 SP - 1-40 SN - 2152-9620 DO - 10.5210/dad.2022.101 UR - https://fis.uni-bamberg.de/handle/uniba/59483 LA - eng ER - TY - CPAPER TI - GraphWOZ: Dialogue Management with Conversational Knowledge Graphs AU - Walker, Nicholas Thomas AU - Ultes, Stefan AU - Lison, Pierre PY - 2022 SP - 1-17 PB - Arxiv.org DO - 10.48550/arXiv.2211.12852 UR - https://fis.uni-bamberg.de/handle/uniba/58420 LA - eng AB - We present a new approach to dialogue management using conversational knowledge graphs as core representation of the dialogue state. To this end, we introduce a new dataset, GraphWOZ, which comprises Wizard-of-Oz dialogues in which human participants interact with a robot acting as a receptionist. In contrast to most existing work on dialogue management, GraphWOZ relies on a dialogue state explicitly represented as a dynamic knowledge graph instead of a fixed set of slots. This graph is composed of a varying number of entities (such as individuals, places, events, utterances and mentions) and relations between them (such as persons being part of a group or attending an event). The graph is then regularly updated on the basis of new observations and system actions. GraphWOZ is released along with detailed manual annotations related to the user intents, system responses, and reference relations occurring in both user and system turns. Based on GraphWOZ, we present experimental results for two dialogue management tasks, namely conversational entity linking and response ranking. For conversational entity linking, we show how to connect utterance mentions to their corresponding entity in the knowledge graph with a neural model relying on a combination of both string and graph-based features. Response ranking is then performed by summarizing the relevant content of the graph into a text, which is concatenated with the dialogue history and employed as input to score possible responses to a given dialogue state. KW - Dialogue Management KW - Knowledge Graph ER -