Revisiting Convolution Architecture in the Realm of DNA Foundation Models
This work addresses the need for efficient and effective DNA foundation models for bioinformatics researchers, though it is incremental as it revisits and enhances an existing architecture.
The paper tackled the problem of comparing classical convolutional networks (CNNs) to newer Transformer and state space model (SSM) architectures in DNA foundation models, and found that their CNN-based method, ConvNova, outperforms recent methods on more than half of tasks, with an average improvement of 5.8% in histone-related tasks while using fewer parameters and enabling faster computation.
In recent years, a variety of methods based on Transformer and state space model (SSM) architectures have been proposed, advancing foundational DNA language models. However, there is a lack of comparison between these recent approaches and the classical architecture convolutional networks (CNNs) on foundation model benchmarks. This raises the question: are CNNs truly being surpassed by these recent approaches based on transformer and SSM architectures? In this paper, we develop a simple but well-designed CNN-based method termed ConvNova. ConvNova identifies and proposes three effective designs: 1) dilated convolutions, 2) gated convolutions, and 3) a dual-branch framework for gating mechanisms. Through extensive empirical experiments, we demonstrate that ConvNova significantly outperforms recent methods on more than half of the tasks across several foundation model benchmarks. For example, in histone-related tasks, ConvNova exceeds the second-best method by an average of 5.8%, while generally utilizing fewer parameters and enabling faster computation. In addition, the experiments observed findings that may be related to biological characteristics. This indicates that CNNs are still a strong competitor compared to Transformers and SSMs. We anticipate that this work will spark renewed interest in CNN-based methods for DNA foundation models.