Phase Change Synaptic Devices for STDP in Neuromorphic Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current neuromorphic systems lack the ability to effectively mimic biological systems' integration of spatiotemporal patterns and extraction of relevant information from noisy inputs, particularly in implementing spike-timing dependent plasticity (STDP) for learning rules.
Innovation Solution
A neuromorphic network utilizing electronic neurons interconnected via Phase Change Memory (PCM) synaptic devices, with a timing controller generating phased operation signals to implement STDP, allowing synaptic conductance to change based on the relative spike times of pre-synaptic and post-synaptic neurons, and incorporating binary state memory devices for probabilistic asynchronous operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If traditional digital models are used for neuromorphic systems, then implementation is straightforward with 0s and 1s, but the ability to mimic biological systems' integration of spatiotemporal patterns and extraction of relevant information from noisy inputs is lost
Solution Approach 1:
The patent replaces traditional digital electronic systems with a neuromorphic system that uses phase-change materials to simulate biological neural network behavior. The phase-change synaptic devices mimic biological synapses by changing their electrical conductance state through phase transitions, enabling the system to process spatiotemporal patterns and implement STDP learning rules like biological systems, while maintaining manufacturability through standard semiconductor fabrication processes.
2Adaptability or versatility
If phase change synaptic devices are used to implement STDP, then synaptic conductance can change based on relative spike times, but the device complexity increases with multiple components per synapse
Solution Approach 1:
The patent merges multiple functions into a single phase-change synaptic device structure. Each synapse comprises a phase-change material layer integrated with electrode structures that together perform both signal transmission and plasticity functions. This consolidation reduces device complexity compared to using separate components for each function while maintaining full STDP implementation capability through the phase-change material's ability to modify conductance based on spike timing.
Solution Approach 2:
The patent utilizes parameter changes in the phase-change material to implement STDP. By controlling the duration, amplitude, and timing of electrical pulses applied to the phase-change synaptic device, the system dynamically adjusts the conductance state to reflect spike timing relationships. This parameter-based control enables complex STDP behavior without requiring structurally complex devices.
3Productivity
If high-density electronic spiking neuronal networks are created, then information integration and processing capabilities are enhanced, but the difficulty of detecting and measuring synaptic states increases
Solution Approach 1:
The patent incorporates feedback mechanisms that enable automatic readout and measurement of synaptic states. The phase-change synaptic devices are designed with integrated sensing capabilities that provide feedback on their conductance states, allowing the system to monitor and measure synaptic weights even in high-density configurations. This feedback approach simplifies detection compared to external measurement methods.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables efficient and phased programming of synaptic weights, mimicking biological STDP learning rules, allowing for high-density electronic spiking neuronal networks that can rapidly extract signals from noisy inputs, enhancing information integration and processing capabilities.
Implementation Method 1
Each synaptic device includes a variable state resistor and a transistor device... the variable state resistor changes state based on phased operation signals... implementing spike-timing dependent plasticity
Data Source
AI summary
One embodiment relates to a neuromorphic network including electronic neurons and an interconnect circuit for interconnecting the neurons. The interconnect circuit includes synaptic devices for interconnecting the neurons via axon paths, dendrite paths and membrane paths. Each synaptic device includes a variable state resistor and a transistor device with a gate terminal, a source terminal and a drain terminal, wherein the drain terminal is connected in series with a first terminal of the variable state resistor. The source terminal of the transistor device is connected to an axon path, the gate terminal of the transistor device is connected to a membrane path and a second terminal of the variable state resistor is connected to a dendrite path, such that each synaptic device is coupled between a first axon path and a first dendrite path, and between a first membrane path and said first dendrite path.