Sarcasm is so easy to bring via schoolbook . Nothing ’s easier . If this sarcastic statement was n’t patent , then perhaps there ’s a reason why researchers are trying to develop new algorithm to better detect it .
Even in person , some sarcastic verbal statements can be dodgy to beak up on . That ’s becausesarcasmis used to convey something other than the literal meaning of the sentence – in most display case , the signification of the sarcastic statement is the contrary of what has actually been say .
If we detect it difficult to always observe sarcasm or sarcasm , computers have an even harder metre . The subtle changes in tone that convey it jumble computing equipment algorithms , which limits virtual assistants and capacity analysis tools .
Now a team of researchers have developed a multimodal algorithmic program that is better equipped to share with our ironical or sarcastic turn of phrase . It does so by see multiple aspects of audio recordings that help oneself it hone its accuracy .
Typically , satire detection algorithms swear on a single parameter to produce their results , which is the main cause why they often miss the mark . To overcome this , the researchers used two complemental approaches – aroused acknowledgement using audio recordings and sentiment analysis that relies on text . Together , these approaches provide a more consummate mental picture of what is going on .
“ We pull out acoustical parameters such as tar , speak rate , and get-up-and-go from voice communication , then used Automatic Speech Recognition to transcribe the delivery into textbook for sentiment analysis , ” Xiyuan Gao of the Speech Technology Lab at the University of Groningen , Campus Fryslân say in astatement .
“ Next , we impute emoticons to each speech segment , chew over its emotional content . By integrate these multimodal cue into a machine eruditeness algorithm , our approach leverages the combined strong point of auditory and textual info along with emoticons for a comprehensive analysis . ”
You may be imagine “ yeah , this is so useful ” ( sarcasm or serious-mindedness ? ) , but the team believes the algorithm has economic value beyond simply identifying wit .
“ The ontogeny of irony realisation technology can benefit other research domains using view analysis and emotion realization , ” Gao add . “ Traditionally , sentiment analytic thinking mainly focuses on text and is train for applications such as online hate delivery sleuthing and client opinion excavation . Emotion recognition ground on voice communication can be apply to AI - assisted health care . Sarcasm recognition engineering science that applies a multimodal approach is insightful to these research arena . ”
Gao and the rest of the team are surefooted that their algorithm will perform well , but they are already exploring way to meliorate it .
“ There are a range of expressions and gesture people apply to highlight sarcastic elements in speech , ” tell Gao . “ These need to be better integrated into our project . In addition , we would care to let in more language and adopt developing irony acknowledgement techniques . ”
If the squad ’s algorithm bring , perhaps one 24-hour interval we will all be free to confidently express our satire without fear we will be see as unfeigned . :P
The work was presented on Thursday , May 16 as part of ajoint meetingof the Acoustical Society of America and the Canadian Acoustical Association .
