This month interesting research appeared on the website of the scientific journal ACL Web. The paper describes a new AI capable of detecting sarcasm on Twitter and Reddit. AI compares a picture and its description, first tests showed 86% accuracy. It аlsо соmраrеs hаshtаgs tо twееt tехt tо hеlр assess thе sentiment а usеr is trying tо cоnvеy.
In modern society, sarcasm became a conventional form of humor, especially among youth. It’s commonly used in memes and sometimes we don’t really understand sarcastic subtext at all. Maybe 5 individuals that created this AI had the same problems and decided to become Kings of the Hill. However, their AI has the potential to become a really promising business idea. An application, based on it, could be, for example, used to analyze comments on the Internet and sort out only serious ones.
As inventors say: “Sarcasm is a pervasive phenomenon in today’s social media platforms such as Twitter and Reddit. These platforms allow users to create multi-modal messages, including texts, images, and videos. Existing multi-modal sarcasm detection methods either simply concatenate the features from multi modalities or fuse the multi modalities information in a designed manner. However, they ignore the incongruity character in sarcastic utterance, which is often manifested between modalities or within modalities. Inspired by this, we propose a BERT architecture-based model, which concentrates on both intra and inter-modality incongruity for multi-modal sarcasm detection.”
It is also crucial to mention that the accuracy of this AI increased compared to a similar one created last year is higher on 2.74%. But the main question is “Why is it useful?”. Well, in my opinion, it’s a little step forward, and using this technology further researches can be conducted. Needless to say that AI usage could be expanded to minimize human work. Like Facebook did the previous year. They used AI to detect inappropriate comments and posts to delete them. I’m sure the same way is possible to get rid out of spam and fishing websites or at least reduce their number. Or instead of sarcasm developers will learn AI to detect inappropriate subtext. Humor and sarcasm are just a part of human psychology and AI keeps exploring us more and more…