Rethinking the bert-like pretraining for dna sequences

chaoqi liang, Weiqiang Bai, Lifeng Qiao, Yuchen Ren, Jianle Sun, Peng Ye, Hongliang Yan, Xinzhu Ma, Wangmeng Zuo, Wanli Ouyang

Primary Area: applications to physical sciences (physics, chemistry, biology, etc.)

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Keywords: Pretrained, DNA, Large Language Model

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.

Abstract:

With the success of large-scale pretraining in NLP, there is an increasing trend of applying it to the domain of life sciences. In particular, pretraining methods based on DNA sequences have garnered growing attention due to their potential to capture generic information about genes. However, existing pretraining methods for DNA sequences largely rely on direct adoptions of BERT pretraining from NLP, lacking a comprehensive understanding and a specifically tailored approach. To address this research gap, we first conducted a series of exploratory experiments and gained several insightful observations: 1) In the fine-tuning phase of downstream tasks, when using K-mer overlapping tokenization instead of K-mer non-overlapping tokenization, both overlapping and non-overlapping pretraining weights show consistent performance improvement.
2) During the pre-training process, using K-mer overlapping tokenization quickly produces clear K-mer embeddings and reduces the loss to a very low level, while using K-mer non-overlapping tokenization results in less distinct embeddings and continuously decreases the loss. 3) Using overlapping tokenization causes the self-attention in the intermediate layers of pre-trained models to tend to overly focus on certain tokens, reflecting that these layers are not adequately optimized. In summary, overlapping tokenization can benefit the fine-tuning of downstream tasks but leads to inadequate pretraining with fast convergence. To unleash the pretraining potential, we introduce a novel approach called RandomMask, which gradually increases the task difficulty of BERT-like pretraining by continuously expanding its mask boundary, forcing the model to learn more knowledge. RandomMask is simple but effective, achieving top-tier performance across 26 datasets spanning 7 downstream tasks. For example, RandomMask achieves a staggering 65.83\% in Matthew's correlation coefficient for epigenetic mark prediction, which is a groundbreaking increase of 14.02\% over the baseline and a remarkable 4.82\% improvement over the SOTA results.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 2654