Computer researchers at the University of Central Florida have developed a sarcasm detector.
Social media has become a dominant form of communication for individuals and for businesses looking to market and sell their products and services. Understanding and responding to customer feedback on Twitter, Facebook and other social media platforms is essential for success, but it takes a lot of work.
This is where sentiment analysis comes in. The term refers to the automated process of identifying the emotion – positive, negative or neutral – associated with the text. While artificial intelligence refers to the logical analysis and response of data, sentiment analysis is akin to the correct identification of emotional communication. A UCF team has developed a technique that accurately detects sarcasm in social media text.
The team’s findings were recently published in the journal Entropy.
In fact, the team taught the computer model to find patterns that often indicate sarcasm, and combined this with teaching the program to correctly select the cue words in sequences that were more likely to indicate sarcasm. They taught the model how to do this by providing it with large data sets and then checking its accuracy.
“The presence of sarcasm in the text is the main obstacle to carrying out sentiment analysis,” says assistant professor of engineering Ivan Garibay ’00MS’ 04PhD. “Sarcasm is not always easy to identify in a conversation, so you can imagine that it is quite difficult for a computer program to do it and do it well. We have developed an interpretable deep learning model using multi-head self-attention and closed recurring units. The multi-head self-attention module helps identify critical sarcastic cues from the entry, and recurring units learn the long-range dependencies between these cues to better categorize the input text. “
The team, which includes PhD student in computer science Ramya Akula, began working on this problem with a DARPA grant that supports the organization’s online social behavior computer simulation program.
“Sarcasm has been a major obstacle to increasing the accuracy of sentiment analysis, especially on social media, as sarcasm relies heavily on vocal sounds, facial expressions, and gestures that cannot be represented in text, ”says Brian Kettler, program manager in DARPA’s Information Office of Innovation (I2O). “Recognizing the sarcasm in online text communication is no easy task because none of these clues are readily available.”
This is one of the challenges that Garibay’s Complex Adaptive Systems Laboratory (CASL) is studying. CASL is an interdisciplinary research group dedicated to the study of complex phenomena such as the global economy, the global information environment, innovation ecosystems, sustainability, and social and cultural dynamics and evolution. . CASLPA scientists study these issues using data science, network science, complexity science, cognitive science, machine learning, deep learning, social science, team cognition , among other approaches.
“In a face-to-face conversation, sarcasm can be identified effortlessly using the speaker’s facial expressions, gestures and tone,” Akula explains. “Detecting sarcasm in text communication is no small task as none of these clues are readily available. Particularly with the explosion in internet use, the detection of sarcasm in online communications at starting from social media platforms is much more difficult. “
Garibay is an assistant professor in industrial engineering and management systems. He has several diplomas including a doctorate. in computer science from UCF. Garibay is the director of the CASL UCF Artificial Intelligence and Big Data Initiative and the master’s program in data analysis. His research areas include complex systems, agent-based models, the dynamics of social media information and disinformation, artificial intelligence and machine learning. It has over 75 peer-reviewed articles and over $ 9.5 million in funding from various national agencies.
Akula is a doctoral candidate and graduate research assistant at CASLPA. She holds an MA in Computer Science from Kaiserslautern Technical University in Germany and a BS in Computer Engineering from Jawaharlal Nehru University of Technology in India.
Source of the story:
Material provided by University of Central Florida. Original written by Zenaida Gonzalez Kotala. Note: Content can be changed for style and length.