Scalable Deep Learning for RNA Secondary Structure Prediction
This work addresses RNA secondary structure prediction, a domain-specific problem in computational biology, with incremental improvements through a novel architecture.
The authors tackled RNA secondary structure prediction by introducing RNAformer, a lean deep learning model using axial attention and recycling in the latent space, achieving state-of-the-art performance on the TS0 benchmark dataset and outperforming methods with external information.
The field of RNA secondary structure prediction has made significant progress with the adoption of deep learning techniques. In this work, we present the RNAformer, a lean deep learning model using axial attention and recycling in the latent space. We gain performance improvements by designing the architecture for modeling the adjacency matrix directly in the latent space and by scaling the size of the model. Our approach achieves state-of-the-art performance on the popular TS0 benchmark dataset and even outperforms methods that use external information. Further, we show experimentally that the RNAformer can learn a biophysical model of the RNA folding process.