RGCL are pleased to announce that one of our PhD students – Hadeel Saadany – has had two papers accepted at COLING 2020 workshops. Congratulations! We look forward to the conference in December.

First paper:

‘Is it Great or Terrible? Preserving Sentiment in Neural Machine Translation of Arabic Reviews’

Accepted in: The Fifth Arabic Natural Language Processing Workshop – COLING’2020, Barcelona, Spain, 12 Dec. 2020

Team: Hadeel Saadany and Constantin Orasan

Summary:

The paper investigates the challenges involved in translating User Generated Content from Arabic into English with particular focus on the errors that lead to incorrect translation of sentiment polarity. It shows that fine-tune an NMT model with respect to sentiment polarity can significantly help in correcting sentiment errors detected in the online translation of Arabic UGC.

Second paper:

‘Fake or Real? A Study of Arabic Satirical Fake News

Accepted in: 3rd International Workshop on Rumours and Deception in Social Media – COLING’2020, Barcelona, Spain, 12 Dec. 2020

Team: Hadeel Saadany, Emad Mohamed, and Constantin Orasan.

 Summary:

This paper conducts several exploratory analyses to identify the linguistic properties of Arabic fake news with satirical content. It shows that although it parodies real news, Arabic satirical news has distinguishing features on the lexico-grammatical level. It also builds a number of machine learning models capable of capturing satirical fake news with an accuracy of up to 98.6\%.