Supervised learning in an adaptive DNA strand displacement circuit

Abstract

The development of DNA circuits capable of adaptive behavior is a key goal in DNA computing, as such systems would have potential applications in long-term monitoring and control of biological and chemical systems. In this paper, we present a framework for adaptive DNA circuits using buffered strand displacement gates, and demonstrate that this framework can implement supervised learning of linear functions. This work highlights the potential of buffered strand displacement as a powerful architecture for implementing adaptive molecular systems.