Nanoscale Resonance Energy Transfer-Based Devices for Probabilistic Computing
Nanoscale Resonance Energy Transfer-Based Devices for Probabilistic Computing
复制标题
用于概率计算的基于纳米共振能量转移的设备
作者：
Siyang Wang;A. Lebeck;C. Dwyer
尽管概率计算在理论上取得了进步，但传统计算机使用的确定性硬件与概率算法的随机性质之间仍然存在根本性的不匹配。在本文中，作者提出发色团之间的共振能量转移（RET）作为概率计算功能单元的支持技术。作为自然的连续时间马尔可夫链，RET网络可以在物理上实现具有任意概率分布的高效采样器，并且具有加速概率算法的巨大潜力。
Despite the theoretical advances in probabilistic computing, a fundamental mismatch persists between the deterministic hardware that traditional computers use and the stochastic nature of probabilistic algorithms. In this article, the authors propose Resonance Energy Transfer (RET) between chromophores as an enabling technology for probabilistic computing functional units. As a natural continuous-time Markov chain, RET networks can physically implement efficient samplers with arbitrary probability distributions and have great potential for accelerating probabilistic algorithms.