Call for papers
CHum 2025 invites submissions of long and short papers on original, unpublished work on the computational processing of humor. All accepted submissions will appear in the published proceedings and be presented as talks at the virtual workshop.
Scope and topics
CHum 2025 aims to foster further work on modeling the processes of humor with current methods in computational linguistics and natural language processing, against the theoretical backdrop of humor research and with reference to relevant corpora of textual, visual, and multimodal materials. A principal goal of the workshop is to unite researchers who can together probe the limits of various meaning representations—symbolic, neural, and hybrid—for humor processing.
We welcome contributions on any topic relevant to the computational processing of humor, including but not limited to the following. Particularly encouraged are submissions describing inter- or multi-disciplinary work, whether completed or in progress, and position papers that critically discuss the past, present, and future of computational humor systems.
- LLMs, knowledge representation
- Using knowledge graphs or symbolic knowledge for humor detection, creation, or appreciation; using LLMs or symbolic knowledge to interpret the reasoning mechanism of joke types; distilling and distinguishing implicit and explicit symbolic knowledge from LLMs needed for humor processing; systematic description of the gaps in LLMs, symbolic, or neurosymbolic approaches for humor processing; new unified representations for symbolic and neural meaning representation, from multidisciplinary humor perspective, in particular for humor generation.
- Resources and evaluation
- Datasets (particularly multimodal ones) useful for training, fine-tuning, or testing general-purpose or task-specific humor models and algorithms; evaluation methodologies and metrics for particular humor tasks or humor types.
- Human–computer interaction
- User interfaces for humor-aware software such as games or (possibly embodied) computational conversational agents, where models or algorithms must detect and classify humor in the human input and react appropriately or generate/recall reciprocating humor and inject it into the interaction.
- Computer-mediated communication
- Similar to the previous area, but where the interactions are between humans, with the computer serving as the medium or facilitator.
- Assisted content creation
- Systems for (interactively) suggesting humorous additions or variants to human-generated semiotic content (text, images, sound, video, or their combination).
- Machine and computer-assisted translation
- Systems to produce, or assist humans in the production of, translation of texts containing humor.
- Digital humanities applications
- (Semi-)automatic methods for assisting humanities scholars with the retrieval, identification, classification, and analysis of humor in literature, music, film, or other (possibly multimodal) media.
- Formal modeling of humor
- Computational resources, models, and methods that allow researchers in the arts and (social) sciences to formalize and test their theories of humor.
- Proof-of-concept humor detection and classification
- Focused academic proofs of concept of certain algorithms and resources against a humor-related benchmark; such fundamental research serving to provide useful building blocks for the above-noted downstream tasks.
Submission instructions
Papers should be formatted according to the same guidelines for the main COLING 2025 conference papers and submitted online through the CHum 2025 submission site on START: https://softconf.com/coling2025/CompHum25/.
Important dates
All deadlines are at 23:59 UTC-12:00 ("anywhere on Earth").
| Initial submission | November 15, 2024 |
|---|---|
| Notification of acceptance | December 2, 2024 |
| Camera-ready submission | December 13, 2024 |
| Workshop | January 19, 2025 |
Contact
To reach the organizers with any questions or comments, please e-mail chum@groups.io.