Artificial Intelligence Can Now Detect Sarcasm. But For What?

This latest advancement of artificial intelligence can be used for national security.

Artificial Intelligence is one step closer to being more human-like as it can now detect sarcasm. Funded by the U.S military, a new AI tool has managed to do a task that is tough for computer algorithms to perform in general, identifying the tone and irony of the human voice. This advancement can help intelligence agencies perform better trend analysis by identifying social media posts that are basically sarcastic in nature and meaning no harm.

How did the AI tool figure it out? According to two researchers from the University of Central Florida, some words in a set of combinations can be a clear indicator of sarcasm in social media posts. Even without any context, it's easy for AI to learn those words and detect sarcasm. The AI tool trains itself with a variety of data from Twitter posts, Reddit threads, dialogues, and media headlines. Garibay and his co-worker Ramya Akula from University listed out the connection between words and sarcastic tones.

"For instance, words such as 'just', 'again', 'totally', and '!' have darker edges connecting them with every other word in a sentence. These are the words in the sentence that hint at sarcasm and as expected, these receive higher attention than others."

Researchers termed this method as self-attention architecture, where complex artificial intelligence programs, known as neural networks, are trained to give more attention to some words than others, depending on what the adjoining words are and what the task at hand is. This might seem silly enough for military importance but considering the social media habits of people right now (which is more than ever), such tools can help officials understand what's happening in key areas where they operate.

Ivan Garibay told Defense One, "Attention is a mechanism to discover patterns in the input that are crucial for solving the given task. In deep learning, self-attention is an attention mechanism for sequences, which helps learn the task-specific relationship between different elements of a given sequence to produce a better sequence representation." It's important to note that this is not the first time this concept is being talked about. This originally dates back to a 2016 research paper written by a German and a Canadian researcher.

The work of this AI tool is supported by the Defence Advanced Research Projects Agency or DARPA via a program called Computational Simulation of Online Social Behavior which aims towards a thorough understanding of adversaries using the global information environment that is possible using this approach.

While this wasn't the first time researchers tried to make a computer algorithm that can detect sarcasm, the current AI tool is an improvement on previous methods. According to Garibay, previous methods used Garibay neural networks to find hidden relationships. But it's still a mystery how the neural networks reached the conclusions. The improvement between the current method and previous neural network methods is the fact that users can now go back and see how the models got the result which is essential in order to use such tools for national security.