SarcasmDet at SemEval-2021 Task 7: Detect Humor and Offensive based on Demographic Factors using RoBERTa Pre-trained Model

Abstract

This paper presents one of the top winning solution systems for task 7 at SemEval2021, "Ha-Hackathon: Detecting and Rating Humor and Offense". The shared task 7 consists of two parts, task-1 with three sub-tasks 1a,1b, and 1c, and task-2. The goal of task-1 is to predict if the text would be considered humorous or not, then if it is yes, predict how humorous it is and whether the humor rating would be perceived as controversial. The goal of task-2 is to predict how the text is considered offensive for users in general. The proposed solution, Sar-casmDet, has been developed using RoBERTa pre-trained model with ensemble techniques. The paper describes the submitted system's architecture with the experiments and the hyperparameter fine-tuning that led to this robust system. Our model ranked third and fourth places out of 50 teams in tasks 1c and 1a with F1-Score of 0.6270 and 0.9675, respectively. At the same time, the model ranked one of the top 10 models in task 1b and task 2 with RMSE scores of 0.5446 and 0.4469, respectively.

Sentiment analysis in SemEval: a review of sentiment identification approaches

“…New PLMs were introduced that outperform existing models to optimize existing models in various directions such as providing lite versions that reduce parameters, increase models speed, and reduce memory consumption (i.e., ALBERT [92] for BERT). Other alternatives focused on training ❒ ISSN: 2088-8708 strategies such as XLNet [93] which introduced an automatic regressive pre-training method [94]. This notable advancement in performance using transformers comes with many limitations including a negative impact of class imbalance [95], and the required intensive computation using large PLM models [96].…”

AraBERT transformer model for Arabic comments and reviews analysis

“…In the next year, Faraj and Abdullah [19] published the best solution for the shared task on sentiment and sarcasm detection in the Arabic language. The objective global of the task is to identify whether a tweet is sarcastic or not.…”

YMAI at SemEval-2022 Task 5: Detecting Misogyny in Memes using VisualBERT and MMBT MultiModal Pre-trained Models

“…Several researchers are attracted to detecting hate, harm, sarcasm, and satire on Social media (Faraj and Abdullah, 2021;Isaksen and Gambäck, 2020;Watanabe et al, 2018). Internet Memes are considered one of the most popular ways to communicate on all topics on social media.…”