Computer ScienceProceedings of the 31st International Conference on Computational Linguistics
Detecting Emotional Incongruity of Sarcasm by Commonsense Reasoning
Z. Qiu, J. Yu, et al.
Leveraging commonsense augmentation, EICR detects sarcasm by using retrieval-augmented large language models to supply missing background, refining dependency graphs to capture contextual associations, and applying an adaptive reasoning skeleton to extract sentiment-inconsistent subgraphs—plus adversarial contrastive learning for robustness. Experiments on five datasets show its effectiveness. Research conducted by Ziqi Qiu, Jianxing Yu, Yufeng Zhang, Hanjiang Lai, Yanghui Rao, Qinliang Su, and Jian Yin.
Related Publications
Explore these studies to deepen your understanding
Adjacent work that informs or extends this paper's methodology and findings.
Physics
4D-imaging of drip-line radioactivity by detecting proton emission from $^{54m}$Ni pictured with ACTAR TPC
J. Giovinazzo, T. Roger, et al.
Social Work
Affective agenda dynamics on social media: interactions of emotional content posted by the public, government, and media during the COVID-19 pandemic
S. Zhou, X. Yang, et al.
Chemistry
Effective screening of Coulomb repulsions in water accelerates reactions of like-charged compounds by orders of magnitude
A. Kowalski, K. Bielec, et al.
Psychology
Impact of technical reasoning and theory of mind on cumulative technological culture: insights from a model of micro-societies
A. Bluet, F. Osiurak, et al.
