Facebook’s Natural Language Processing (NLP) algorithms are failing to detect whether its users are being sarcastic or ironic in their comments and likes, according to a leading social media expert.
According to Forbes, social media expert Matthew Russell, author of Mining the Social Web, has called into question whether the NLP algorithms used by Facebook are able to accurately detect when its users genuinely “like” a product and when they are being ironic.
Russell said “context matters. Analysing natural language data is, in my opinion, the problem of the next two to three decades.”
“It’s an incredibly difficult issue, and sarcasm and other types of ironic language are inherently problematic for machines to detect when looked at in isolation. It’s imperative to have a sufficiently sophisticated and rigorous enough approach that relevant context can be taken into account.”
For potential advertisers, the value of advertising with Facebook derives in large part from its ability to use its large user database to accurately target potential consumers.
In turn, business confidence in Facebook’s ability to deliver highly targeted advertising is essential to the company monetising its large user base following its recent IPO.
Andrew Sadauskas is a former journalist at SmartCompany and a former editor of TechCompany.
