AI can detect sarcasm now. Great...
Reading the room perfectly for how we feel about artificial intelligence (AI), the Speech Technology Lab at the University of Groningen has finally created a programme able to discern tone of voice and sarcasm.
It might be the lowest form of wit, but the AI is an impressive development through displaying how technology can understand more subjective aspects of text.
In a presentation to the Acoustical Society of America and the Canadian Acoustical Association, the researchers explained how the prevalence of sarcasm in human-written text means that without AI properly detecting it, there would be huge gaps in the potential capabilities of AI text comprehension.
Previous AI sarcasm detectors only used one parameter, but this new technology combines methods. “We extracted acoustic parameters such as pitch, speaking rate, and energy from speech, then used Automatic Speech Recognition to transcribe the speech into text for sentiment analysis,” said researcher Xiyuan Gao.
“Next, we assigned emoticons to each speech segment, reflecting its emotional content. By integrating these multimodal cues into a machine learning algorithm, our approach leverages the combined strengths of auditory and textual information along with emoticons for a comprehensive analysis,” Gao explained.
The team relied on the Mustard (Multimodal Sarcasm Detection Dataset) database of sarcasm examples in popular TV shows to teach their AI about sarcasm through clips from ‘The Big Bang Theory’ and ‘Friends’.
“There are a range of expressions and gestures people use to highlight sarcastic elements in speech,” said Gao. “These need to be better integrated into our project. In addition, we would like to include more languages and adopt developing sarcasm recognition techniques.”
Tone of language is one of the more complex elements of text. Many people find it difficult to detect sarcasm when it’s written down without any audio/visual cues. Not you though, you’re far too smart for that, aren’t you?
Another researcher on the team, Shekhar Nayak, has suggested that while the project will have benefits for improving human-AI text interactions, it also could have applications in detecting abuse and hate speech.
“The development of sarcasm recognition technology can benefit other research domains using sentiment analysis and emotion recognition,” said Gao.
She continued: “Traditionally, sentiment analysis mainly focuses on text and is developed for applications such as online hate speech detection and customer opinion mining. Emotion recognition based on speech can be applied to AI-assisted health care. Sarcasm recognition technology that applies a multimodal approach is insightful to these research domains.”
