MIT software shows how NLP systems are snookered by simple synonyms
Here’s an example of how artificial intelligence can still seriously lack behind some human attributes: tests have shown how natural language processing (NLP) systems can be tricked into misunderstanding text by merely swapping one word for a synonym.
A research team at MIT developed software, called TextFooler, which looked for words which were most crucial to an NLP classifier and replaced them. The team offered an example:
“The characters, cast in impossibly...