Sarcasm is so easy to convey via text. Nothing’s easier. If this sarcastic statement wasn’t apparent, then perhaps there’s a reason why researchers are trying to develop new algorithms to better detect it.
Even in person, some sarcastic verbal statements can be tricky to pick up on. That’s because sarcasm is used to convey something other than the literal meaning of the sentence – in most cases, the meaning of the sarcastic statement is the opposite of what has actually been said.
If we find it difficult to always detect sarcasm or irony, computers have an even harder time. The subtle changes in tone that convey it confuse computer algorithms, which limits virtual assistants and content analysis tools.
Now a team of researchers have developed a multimodal algorithm that is better equipped to deal with our ironic or sarcastic turns of phrase. It does so by examining multiple aspects of audio recordings that help it hone its accuracy.
Typically, sarcasm detection algorithms rely on a single parameter to produce their results, which is the main reason why they often miss the mark. To overcome this, the researchers used two complementary approaches – emotional recognition using audio recordings and sentiment analysis that relies on text. Together, these approaches provide a more complete picture of what is going on.
“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,” Xiyuan Gao of the Speech Technology Lab at the University of Groningen, Campus Fryslân said in a statement.
“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.”
You may be thinking “yeah, this is so useful” (sarcasm or sincerity?), but the team believes the algorithm has value beyond simply identifying wit.
“The development of sarcasm recognition technology can benefit other research domains using sentiment analysis and emotion recognition,” Gao added. “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.”
Gao and the rest of the team are confident that their algorithm will perform well, but they are already exploring ways to improve it.
“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.”
If the team’s algorithm works, perhaps one day we will all be free to confidently express our sarcasm without fear we will be seen as genuine. 😛
The work was presented on Thursday, May 16 as part of a joint meeting of the Acoustical Society of America and the Canadian Acoustical Association.
Source Link: An Algorithm To Detect Sarcasm? Like We Need That
