Israeli Machine Learning System Identifies Sarcasm, Irony and Emotions By Scanning Social Networks
Israeli computer science engineer Eden Saig from the Technion University has developed a new machine learning system that can classify sentiments 'hidden' in the text shared on social networks. Saig has been a student of artificial intelligence (AI) and has published a paper titled Sentiment Classification of Texts in Social Networks, under guidance of Prof. Shaul Markovich. The new AI system is intelligent enough to figure out sarcasm, iron and emotions in the text messages.
Saig's system works by identifying word patterns that repeat in the texts. In verbal communication, the tone and vocal inflections carry a lot of information about the intentions and feelings; but these two crucial factors aren't available in textual communication. The use of smiley emoticons accompanied with text messages help address this problem to some extent, but they do not always convey the right message. Saig says that these emoticons are very superficial cues.
Saig had to analyse several thousands of posts on Facebook pages in Hebrew. He then isolated the patterns and developed his AI machine learning system to pick up these patterns and figure out the trends across social media.
Engineer Eden Saig. Image Credit: <a href="https://webcourse.cs.technion.ac.il/234900/Spring2015/en/staff.html" target="_blank" rel="noopener noreferrer">234900 - Workshop in Competitive Programming, Spring2015 - Staff</a>
At this point, Saig's system can recognise caring sentiments or arrogant ones. The system can be useful to scan content shared by people on Twitter, Facebook and other popular social networks to detect anti-social elements and even find out persons with suicidal thoughts.
Saig's ultimate aim however is to develop a mechanism that will help people understand how their words can be interpreted by others. He believes that such system will allow people to express themselves better and avoid unnecessary misunderstanding.
Do let us know your thoughts on the topic. Move coverage available on the source link mentioned below.
Source: <a href="https://www.timesofisrael.com/social-media-analysis-may-help-prevent-terror-attacks-study-shows/" target="_blank" rel="noopener noreferrer">Social media analysis may help prevent terror attacks, study shows | The Times of Israel</a>
Saig's system works by identifying word patterns that repeat in the texts. In verbal communication, the tone and vocal inflections carry a lot of information about the intentions and feelings; but these two crucial factors aren't available in textual communication. The use of smiley emoticons accompanied with text messages help address this problem to some extent, but they do not always convey the right message. Saig says that these emoticons are very superficial cues.
Saig had to analyse several thousands of posts on Facebook pages in Hebrew. He then isolated the patterns and developed his AI machine learning system to pick up these patterns and figure out the trends across social media.
Engineer Eden Saig. Image Credit: <a href="https://webcourse.cs.technion.ac.il/234900/Spring2015/en/staff.html" target="_blank" rel="noopener noreferrer">234900 - Workshop in Competitive Programming, Spring2015 - Staff</a>
At this point, Saig's system can recognise caring sentiments or arrogant ones. The system can be useful to scan content shared by people on Twitter, Facebook and other popular social networks to detect anti-social elements and even find out persons with suicidal thoughts.
Saig's ultimate aim however is to develop a mechanism that will help people understand how their words can be interpreted by others. He believes that such system will allow people to express themselves better and avoid unnecessary misunderstanding.
Do let us know your thoughts on the topic. Move coverage available on the source link mentioned below.
Source: <a href="https://www.timesofisrael.com/social-media-analysis-may-help-prevent-terror-attacks-study-shows/" target="_blank" rel="noopener noreferrer">Social media analysis may help prevent terror attacks, study shows | The Times of Israel</a>
0