Closed kharaldsson closed 1 year ago
Affect in Tweets Datasets: https://competitions.codalab.org/competitions/17751#learn_the_details-datasets Evaluation script: https://github.com/felipebravom/SemEval_2018_Task_1_Eval Paper: http://saifmohammad.com/WebDocs/semeval2018-task1.pdf 5 subtasks - a mix of regression and classification Adaptation task ideas: English vs. Arabic (Language), Regression vs. Classification (Prediction Type)
Irony in English Tweets Datasets: https://github.com/Cyvhee/SemEval2018-Task3/tree/master/datasets Evaluation script: https://github.com/Cyvhee/SemEval2018-Task3/tree/master/evaluation Paper: https://aclanthology.org/S18-1005.pdf Adaptation task ideas: Tweets vs. Reviews (Genre), Irony vs. Humor (Affect Type)
Humor in Headlines Datasets: https://www.cs.rochester.edu/u/nhossain/humicroedit.html Evaluation script not provided Paper: https://aclanthology.org/N19-1012.pdf Adaptation task ideas: Regression vs. Classification, Humor vs. Sarcasm, New vs. Tweets
Sarcasm Detection: https://sites.google.com/view/semeval2022-isarcasmeval#h.t53li2ejhrh8 Dataset: https://github.com/iabufarha/iSarcasmEval Evaluation: Prescribed in readme on git repo Paper: https://aclanthology.org/2022.semeval-1.111/ Adaptation: English / Arabic
Hate Speech (HatEval): https://competitions.codalab.org/competitions/19935 Dataset: I got my hands on the data for this via the request form. Have saved it locally if we use it. Paper: https://aclanthology.org/S19-2007/ Evaluation: https://github.com/msang/hateval/tree/master/SemEval2019-Task5/evaluation Adaption: Langauge (english, Spanish); Domain (how does this work on "hyperpartisan news dataset"?)
Humor in Headlines then adapt to tweets (HaHackathon) or vice versa...
Affect in Tweets Task E-c: Detecting Emotions (multi-label classification), or Emotion Intensity
Humor in Headlines
Empathy, Emotion, Personality https://codalab.lisn.upsaclay.fr/competitions/834 https://codalab.lisn.upsaclay.fr/competitions/11167 Dataset
Paper: https://arxiv.org/abs/2205.12698 Evaluation: https://drive.google.com/file/d/1BFBlM7fqd8DqLfon9hAQRz5FvH5yya61/view Adaption: reactions to news articles, and conversations between two users
Decided on HatEval
Hate Speech (HatEval): https://competitions.codalab.org/competitions/19935 Dataset: http://hatespeech.di.unito.it/hateval.html Paper: https://aclanthology.org/S19-2007/ Evaluation: https://github.com/msang/hateval/tree/master/SemEval2019-Task5/evaluation Adaption: Langauge (english, Spanish); Domain (how does this work on "hyperpartisan news dataset"?)
Affect Tasks from Canvas