Advancing Cognitive AI at ICMLANT 2024

Dr. Mageda Sharafeddin, Dean of the College of Arts and Sciences, Mr. Tareq Ayoub, a faculty member from the Social Sciences Department, and Jad Raad, a senior Computer Science student, have co-authored a research article recently presented at the 2024 IEEE International Conference on Machine Learning and Applied Network Technologies (ICMLANT).

The paper explores the evolution from emotion AI to cognitive AI, examining how AI can make informed decisions based on the detection of sentiment data. Focusing specifically on irony detection across various Arabic dialects, the study leverages data from X (formerly Twitter)—a rich and open source of language-structured data. The team utilized advanced models like Neural Networks, BERT, and AraBERT (an Arabic-specific adaptation of BERT) to enhance AI’s ability to classify and interpret sentiment, with a special emphasis on detecting irony. Their models delivered impressive results, with AraBERT surpassing previous benchmarks by achieving an F1-score of 96% on one dataset, in addition to demonstrating strong performance across multiple others.

By advancing beyond sentiment analysis toward broader cognitive capabilities, the study paves the way for future research aimed at shaping the development of cognitive AI systems. Moreover, the collaboration between the Computer Science and Social Sciences departments at PU combined technical expertise in AI with insights into human behavior in digital spaces. Through this interdisciplinary approach, PU continues to drive innovation and advance the development of AI technologies that bridge the gap between technology and human interaction.