Reservoir Computing (RC) refers to a Recurrent Neural Networks (RNNs) framework, frequently used for sequence learning and time series prediction. The RC system consists of a random fixed-weight RNN (the input-hidden reservoir layer) and a classifier (the hidden-output readout layer). Here we focus on the sequence learning problem, and we explore a different approach to RC. More specifically, we remove the non-linear neural activation function, and we consider an orthogonal reservoir acting on normalized states on the unit hypersphere. Surprisingly, our numerical results show that the system's memory capacity exceeds the dimensionality of the reservoir, which is the upper bound for the typical RC approach based on Echo State Networks (ESNs). We also show how the proposed system can be applied to symmetric cryptography problems, and we include a numerical implementation.
Inferring a generative model from data is a fundamental problem in machine learning. It is well-known that the Ising model is the maximum entropy model for binary variables which reproduces the sample mean and pairwise correlations. Learning the parameters of the Ising model from data is the challenge. We establish an analogy between the inverse Ising problem and the Ornstein-Zernike formalism in liquid state physics. Rather than analytically deriving the closure relation, we use a deep neural network to learn the closure from simulations of the Ising model. We show, using simulations as well as biochemical datasets, that the deep neural network model outperforms systematic field-theoretic expansions and can generalize well beyond the parameter regime of the training data. The neural network is able to learn from synthetic data, which can be generated with relative ease, to give accurate predictions on real world datasets.
We show that dispersive propagation of light followed by phase detection has properties that can be exploited for extracting features from the waveforms. This discovery is spearheading development of a new class of physics-inspired algorithms for feature extraction from digital images with unique properties and superior dynamic range compared to conventional algorithms. In certain cases, these algorithms have the potential to be an energy efficient and scalable substitute to synthetically fashioned computational techniques in practice today.