Classification with a disordered dopant-atom network in silicon.

Classification is an important task at which both biological and artificial neural networks excel1,2. In machine learning, nonlinear projection into a high-dimensional feature space can make data linearly separable3,4, simplifying the classification of complex features. Such nonlinear projections are computationally expensive in conventional computers. A promising approach is to exploit physical materials systems that perform this nonlinear projection intrinsically, because of their high computational density5, inherent parallelism and energy efficiency6,7. However, existing approaches either rely on the systems' time dynamics, which requires sequential data processing and therefore hinders parallel computation5,6,8, or employ large materials systems that are difficult to scale up7. Here we use a parallel, nanoscale approach inspired by filters in the brain1 and artificial neural networks2 to perform nonlinear classification and feature extraction. We exploit the nonlinearity of hopping conduction9-11 through an electrically tunable network of boron dopant atoms in silicon, reconfiguring the network through artificial evolution to realize different computational functions. We first solve the canonical two-input binary classification problem, realizing all Boolean logic gates12 up to room temperature, demonstrating nonlinear classification with the nanomaterial system. We then evolve our dopant network to realize feature filters2 that can perform four-input binary classification on the Modified National Institute of Standards and Technology handwritten digit database. Implementation of our material-based filters substantially improves the classification accuracy over that of a linear classifier directly applied to the original data13. Our results establish a paradigm of silicon-based electronics for small-footprint and energy-efficient computation14.

In-Materio Reservoir Computing in a Sulfonated Polyaniline Network.

A sulfonated polyaniline (SPAN) organic electrochemical network device (OEND) is fabricated using a simple drop-casting method on multiple Au electrodes for use in reservoir computing (RC). The SPAN network has humidity-dependent electrical properties. Under high humidity, the SPAN OEND exhibits mainly ionic conduction, including charging of an electric double layer and ionic diffusion. The nonlinearity and hysteresis of the current-voltage characteristics progressively increase with increasing humidity. The rich dynamic output behavior indicates wide variations for each electrode, which improves the RC performance because of the disordered network. For RC, waveform generation and short-term memory tasks are realized by a linear combination of outputs. The waveform task accuracy and memory capacity calculated from a short-term memory task reach 90% and 33.9, respectively. Improved spoken-digit classification is realized with 60% accuracy by only 12 outputs, demonstrating that the SPAN OEND can manage time series dynamic data operation in RC owing to a combination of rich dynamic and nonlinear electronic properties. The results suggest that SPAN-based electrochemical systems can be applied for material-based computing, by exploiting their intrinsic physicochemical behavior.

A deep-learning approach to realizing functionality in nanoelectronic devices.

Many nanoscale devices require precise optimization to function. Tuning them to the desired operation regime becomes increasingly difficult and time-consuming when the number of terminals and couplings grows. Imperfections and device-to-device variations hinder optimization that uses physics-based models. Deep neural networks (DNNs) can model various complex physical phenomena but, so far, are mainly used as predictive tools. Here, we propose a generic deep-learning approach to efficiently optimize complex, multi-terminal nanoelectronic devices for desired functionality. We demonstrate our approach for realizing functionality in a disordered network of dopant atoms in silicon. We model the input-output characteristics of the device with a DNN, and subsequently optimize control parameters in the DNN model through gradient descent to realize various classification tasks. When the corresponding control settings are applied to the physical device, the resulting functionality is as predicted by the DNN model. We expect our approach to contribute to fast, in situ optimization of complex (quantum) nanoelectronic devices.