Out of Order: How Important Is The Sequential Order of Words in a Sentence in Natural Language Understanding Tasks?
This work highlights a limitation in current state-of-the-art NLU models, showing they often don't truly understand sentence meaning, which is critical for researchers and developers aiming to build more robust and human-like AI.
This paper investigates the importance of word order for BERT-based models on various natural language understanding tasks. They found that 75% to 90% of correct predictions on GLUE tasks remain constant even after input words are randomly shuffled, indicating that models often rely on superficial cues rather than sequential order. Encouraging the capture of word order information improves performance across most GLUE tasks, SQuAD 2.0, and out-of-sample data.
Do state-of-the-art natural language understanding models care about word order - one of the most important characteristics of a sequence? Not always! We found 75% to 90% of the correct predictions of BERT-based classifiers, trained on many GLUE tasks, remain constant after input words are randomly shuffled. Despite BERT embeddings are famously contextual, the contribution of each individual word to downstream tasks is almost unchanged even after the word's context is shuffled. BERT-based models are able to exploit superficial cues (e.g. the sentiment of keywords in sentiment analysis; or the word-wise similarity between sequence-pair inputs in natural language inference) to make correct decisions when tokens are arranged in random orders. Encouraging classifiers to capture word order information improves the performance on most GLUE tasks, SQuAD 2.0 and out-of-samples. Our work suggests that many GLUE tasks are not challenging machines to understand the meaning of a sentence.