Neural networks in the brain are structured in three-dimensional (3D) space, and the networks evolve through development and learning, whereas two-dimensional (2D) crossbars have essentially been optimized for a fully connected neural network, which results in a significant increase in unused memristors. Here, we present a prototype of molecular neural networks on wetware consisting of a space-free synaptic medium immersed in monomer solution. In the medium, conductive polymer wires are grown between multiple electrodes through learning only when necessary, i.e. no polymer wire is pre-placed, unlike present 2D crossbar devices. Through experiments, we found the necessary growth conditions for synaptic polymer wires. We first demonstrated the learning of simple Boolean functions and then data-encoding tasks by using our system comprising the synaptic media and their external controllers. These results are valuable for expanding the concept of space-free synapse development, i.e. extending our 2D synaptic media to 3D is possible in principle.

The human brain possesses an exceptional information processing capability owing to the 3D and dense network architecture of numerous neurons and synapses. Brain‐inspired neuromorphic hardware can also benefit from 3D architectures, such as high integration of circuits and acquisition of highly complex dynamical systems. In this study, for future 3D neuromorphic engineering, 3D conductive polymer networks consisting of poly(3,4‐ethylenedioxy‐thiophene) doped with poly(styrene sulfonate) anions (PEDOT:PSS) are successfully and stably fabricated between multiple electrodes from scratch in precursor solution by electropolymerization. The networks efficiently emulate the 3D local connections between neighboring neurons observed in the cortex. This novel technology, which allows 3D conductive wiring only between desired electrodes, is unprecedented and has potential as an underlying technology for 3D integration. Furthermore, the experimental results also conclusively prove that conductance modification can be performed by manipulating the physical and chemical properties of 3D branch‐wired conductive polymer wires, thus demonstrating for the first time the feasibility of neuromorphic wetware with enhanced biological plausibility in the subsequent post‐Moore era.

Networks in the human brain are extremely complex and sophisticated. The abstract model of the human brain has been used in software development, specifically in artificial intelligence. Despite the remarkable outcomes achieved using artificial intelligence, the approach consumes a huge amount of computational resources. A possible solution to this issue is the development of processing circuits that physically resemble an artificial brain, which can offer low-energy loss and high-speed processing. This study demonstrated the synaptic functions of conductive polymer wires linking arbitrary electrodes in solution. By controlling the conductance of the wires, synaptic functions such as long-term potentiation and short-term plasticity were achieved, which are similar to the manner in which a synapse changes the strength of its connections. This novel organic artificial synapse can be used to construct information-processing circuits by wiring from scratch and learning efficiently in response to external stimuli.

Predictive coding (PC) based on the free-energy principle (FEP) has been extensively explored for the next-generations artificial intelligence. In this study, we constructed PC network models that can implement perception and unsupervised learning by minimizing variational free energy through neural dynamics. Furthermore, these models were applied to practical tasks, such as image discrimination and real-time prediction. For implementation as neuromorphic hardware with biological plausibility, the performance of PC networks was evaluated by applying techniques used in recent neuromorphic engineering, such as augmented direct feedback alignment and physical reservoir systems. These results pave the way for neuromorphic hardware capable of autonomous perception and learning based on the FEP.

Recently, there has been considerable research on nonvolatile analog devices for artificial intelligence (AI); however, it focuses on all-coupled neural networks. In contrast, polymer wire-type synaptic devices, which can be expected to be arbitrarily wired similar to a biological neural network, have already been proposed and demonstrated. In this study, we model a polymer wire synaptic device based on the results of previous research, and demonstrate an example of applying simple perceptron (AI) to the model. The results of our study show that it is possible to predict effective methods of using polymer wire synaptic elements in AI.

A hardware neural network utilizing responses from devices, materials, or physical substances for which a function can be used for information processing is now attracting considerable attention. We demonstrated the structures of two different types of neural networks utilizing organic materials and ions. One is a conductive polymer wire, which can crosslink arbitrary electrodes in solution, exhibiting synaptic functions, such as long-term potentiation and short-term plasticity. The other one is electrochemical reactions, which have functionality of an emerging flamework of recurrent neural network, reservoir computing. The results have demonstrated the potential of chemical dynamics as a computing resource. These results will help to construct neuromorphic wetware mimicking a real brain, in the future.