SIGDIAL 2026 (Meeting of the Special Interest Group on Discourse and Dialogue) is a ICORE B conference held in Atlanta, Georgia, USA on 2026-08-02. The paper submission deadline is 2026-04-13. Acceptance notifications are sent on 2026-06-01.
Topics of Interest
SIGDIAL 2026 invites submissions of original research on all aspects of discourse and dialogue. We encourage formal, corpus-based, experimental, or analytical work, as well as work on implementations and applications, including but not limited to the following areas:
Agentic AI and Interaction: Research focused on autonomous agents where dialogue serves as a core mechanism for reasoning, planning, and action execution. This includes contributions on tool-augmented dialogue systems, cognitive architectures for agents (integrating memory, reflection, and goal decomposition), and agentic workflows. Specific areas of interest include multi-agent coordination and negotiation, human-agent teaming and delegation, interactive task learning, and the safety and robustness of agents acting in digital or physical environments. We also encourage submissions on evaluating agentic capabilities, distinguishing between conversational fluency and functional success in complex, multi-step tasks.
Corpora, Tools, and Methodology: Submissions focused on corpus-based and experimental work on discourse and dialogue. We encourage submissions related to annotation tools and schemes, crowdsourcing, evaluation methodologies, and corpora development.
Dialogue Systems: Contributions related to task-oriented and open-domain dialogue systems, whether spoken, multimodal, embedded, situated, or text-based. This includes research on system components, evaluation, and applications. Specific areas of interest include knowledge representation and extraction for dialogue, state representation and tracking, policy learning, social and emotional intelligence, dialogue in virtual reality and human-robot interaction, entrainment, alignment, and priming. We also welcome work on generation for dialogue, style, voice, personality, and safety and ethics in dialogue systems.
Discourse Processing: Research on rhetorical and coherence relations, discourse parsing, discourse connectives, reference resolution, event representation, and causality in narrative. This also includes work on argument mining, text quality and style, cross-lingual discourse analysis, and discourse considerations in applications like machine translation, text summarization, essay grading, question answering, and information retrieval. We particularly encourage submissions that explore discourse issues in text generated by large language models.
LLM-Based Dialogue Technologies: Research on the use of Large Language Models (LLMs) in dialogue systems, including but not limited to areas like prompt engineering, fine-tuning for dialogue, data synthesis and augmentation for dialogue tasks, safety and ethics of LLMs in conversation, and evaluation of LLM-generated dialogue.
Pragmatic and Semantic Modeling: Investigations into the pragmatics and semantics of conversations, going beyond the single sentence level. This includes research on rational speech acts, conversation acts, intentions, conversational implicature, and presuppositions.
User Simulation and Persona Modeling: Research on creating and utilizing computational models of human users to simulate conversational behaviors, preferences, traits, and social dynamics. This includes work on LLM-based user simulators, role-playing agents, and the generation of synthetic interaction data. Specific areas of interest include maintaining persona consistency and memory, modeling diverse user demographics and linguistic styles, and evaluating the fidelity of user simulators. We particularly encourage submissions exploring the use of simulators for offline evaluation of dialogue systems, training strategies (e.g., RLHF, RLAIF), and stress-testing system safety and robustness.
Applications of Dialogue and Discourse Processing: We welcome submissions showcasing innovative applications of dialogue and discourse processing technology in various domains.
まだコメントはありません。