Demonstration of Decentralized Physics-Driven Learning
Demonstration of Decentralized Physics-Driven Learning
复制标题
DOI：
10.1103/physrevapplied.18.014040
复制
发表时间：
2022-07-18
影响因子：
4.6
通讯作者：
Durian, Douglas J.
中科院分区：
文献类型：
--
作者：
Dillavou, Sam;Stern, Menachem;Durian, Douglas J.
在典型的人工神经网络中，神经元根据中央处理器的全局计算进行调整，但在大脑中，神经元和突触基于局部信息进行自我调整。最近提出了对比学习算法来训练物理系统，例如流体、机械或电气网络，以根据局部进化规则执行机器学习任务。然而，到目前为止，这种系统只在计算机上实现，这是由于创建基于其自身对两组全局边界条件的响应而自主进化的元素的工程挑战。在这里，我们介绍并实现了一个物理驱动的对比学习方案的可变电阻器的网络，使用电路来本地比较两个相同的网络受到两组不同的边界条件的响应。使用这种方法，我们的系统可以有效地训练自己，优化其电阻值，而无需使用中央处理器或外部信息存储。一旦系统被训练用于指定的变构、回归或分类任务，则随后通过物理命令快速且自动地执行该任务，以响应于给定的电压输入来最小化功率耗散。我们证明，与典型的计算机不同，这种学习系统由于其分散学习而对极端损坏（以及制造缺陷）具有鲁棒性。因此，我们的双网络方法很容易扩展到非常大或非线性的网络，其分布式特性将是一个巨大的优势;只有500条边的实验室网络已经超过了它的计算机网络。
In typical artificial neural networks, neurons adjust according to global calculations of a central processor, but in the brain, neurons and synapses self-adjust based on local information. Contrastive learning algorithms have recently been proposed to train physical systems, such as fluidic, mechanical, or electrical networks, to perform machine-learning tasks from local evolution rules. However, to date, such systems have only been implemented in silico due to the engineering challenge of creating elements that autonomously evolve based on their own response to two sets of global boundary conditions. Here, we introduce and implement a physics-driven contrastive learning scheme for a network of variable resistors, using circuitry to locally compare the response of two identical networks subjected to the two different sets of boundary conditions. Using this method, our system effectively trains itself, optimizing its resistance values without the use of a central processor or external information storage. Once the system is trained for a specified allostery, regression, or classification task, the task is subsequently performed rapidly and automatically by the physical imperative to minimize power dissipation in response to the given voltage inputs. We demonstrate that, unlike typical computers, such learning systems are robust to extreme damage (and thus manufacturing defects) due to their decentralized learning. Our twin-network approach is therefore readily scalable to extremely large or nonlinear networks, where its distributed nature will be an enormous advantage; a laboratory network of only 500 edges will already outpace its in silico counterpart.