A bioinspired configurable cochlea based on memristors

Abstract

Cochleas are the basis for biology to process and recognize speech information, emulating which with electronic devices helps us construct high-efficient intelligent voice systems. Memristor provides novel physics for performing neuromorphic engineering beyond complementary metal-oxide-semiconductor technology. This work presents an artificial cochlea based on the shallen-key filter model configured with memristors, in which one filter emulates one channel. We first fabricate a memristor with the TiN/HfOx/TaOx/TiN structure to implement such a cochlea and demonstrate the non-volatile multilevel states through electrical operations. Then, we build the shallen-key filter circuit and experimentally demonstrate the frequency-selection function of cochlea’s five channels, whose central frequency is determined by the memristor’s resistance. To further demonstrate the feasibility of the cochlea for system applications, we use it to extract the speech signal features and then combine it with a convolutional neural network to recognize the Free Spoken Digit Dataset. The recognition accuracy reaches 92% with 64 channels, compatible with the traditional 64 Fourier transform transformation points of mel-frequency cepstral coefficients method with 95% recognition accuracy. This work provides a novel strategy for building cochleas, which has a great potential to conduct configurable, high-parallel, and high-efficient auditory systems for neuromorphic robots.

Mimicking biological synapses with a-HfSiOx-based memristor: implications for artificial intelligence and memory applications

“…The pre-synaptic spike onto the pre-synapse is analogous to the applied electric pulse on the TE, which stimulates ion migration via the a-HfSiO x switching layer and causes a change in conductance. Since varying synaptic strength impacts synaptic plasticity, the a-HfSiO x insulating layer mimics the synaptic cleft, allowing neurotransmitters to pass from pre- to post-neuron [ 81 ]. Basically, synaptic plasticity, the mechanism behind signal transmission between two neurons, is a change in synaptic weight in response to environmental inputs.…”

Short-term synaptic plasticity in emerging devices for neuromorphic computing

“…As for ITD, these signals exhibit phase-locked firing actions. 109 When the sound localization is activated, the nervous system acts as an amplifier which would magnify the time difference generated by the ITD twice in the neural circuit. 110 As shown in Figure 6 C, the red line represents the acoustical signal received from the environment.…”

An FPGA-Based Training System for a 1T1R Memristor Array With 500 nS Conductance Resolution Limit

“…For the memristor array utilized in this study [42], [43], a 1T1R cell consists of an NMOS transistor and a memristor, where the NMOS transistor serves as the selector with its drain connected to the top electrode of the memristor. The 1024 1T1R cells are organized into a 32 × 32 array, with 32 word lines connected to the gates of the transistors and the memristor devices positioned between the bit lines and the drain of the transistors.…”