@article(2026-mesironic-a-multimodal-dataset-and-benchmark-for-irony-detection-in-spontaneous-spoken-spanish,
  author = {Vicent Ahuir and Alejandro Joaquín Barceló-Milkova and Andreu Casamayor-Segarra and Valentina González Verdegais and Lluís-F. Hurtado and Carlos Periñán-Pascual and Maria Jose Castro-Bleda},
  title = {MESironic: A Multimodal Dataset and Benchmark for Irony Detection in Spontaneous Spoken Spanish},
  year = {2026},
  journal = {Procesamiento del Lenguaje Natural},
  volume = {77},
  number = {0},
  pages = {347-355},
  abstract = {We introduce MESironic, a multimodal corpus for irony detection in spontaneous Peninsular Spanish. The dataset comprises 6000 short audio clips with manually verified transcriptions, evenly split between ironic and non?ironic instances. Each sample contains the raw waveform and its text transcription. We describe the source material, the annotation protocol, and the main statistical properties of the collection. To provide a first benchmark, we evaluate a text?only baseline (TF?IDF+SVC) and a simple multimodal baseline that concatenates MFCC?derived acoustic features with the same textual representation. The multimodal system outperform the text?only model, confirming that acoustic and linguistic cues are complementary for irony detection. The dataset is publicly available to foster research in multimodal irony understanding in Spanish.}
)