How does the task complexity of masked pretraining objectives affect downstream performance?
This work addresses the optimization of pretraining strategies for NLP models, offering incremental insights into task complexity for researchers and practitioners.
The paper tackled the problem of how the complexity of masked pretraining objectives affects downstream performance, finding that more complex objectives tend to yield better results, with at least half the complexity of masked language modeling needed to match its performance on benchmarks like GLUE, SQuAD, and Universal Dependencies.
Masked language modeling (MLM) is a widely used self-supervised pretraining objective, where a model needs to predict an original token that is replaced with a mask given contexts. Although simpler and computationally efficient pretraining objectives, e.g., predicting the first character of a masked token, have recently shown comparable results to MLM, no objectives with a masking scheme actually outperform it in downstream tasks. Motivated by the assumption that their lack of complexity plays a vital role in the degradation, we validate whether more complex masked objectives can achieve better results and investigate how much complexity they should have to perform comparably to MLM. Our results using GLUE, SQuAD, and Universal Dependencies benchmarks demonstrate that more complicated objectives tend to show better downstream results with at least half of the MLM complexity needed to perform comparably to MLM. Finally, we discuss how we should pretrain a model using a masked objective from the task complexity perspective.