Neuromorphic Memory Element Emulating Neuron and Synapse
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computer systems face bottlenecks and high energy consumption due to data exchange between processing units and memory in the era of big data, necessitating a more efficient neuromorphic computing solution that mimics human brain functionality.
Innovation Solution
A neuromorphic memory element is developed, combining volatile and non-volatile features in a single element to emulate neurons and synapses, utilizing a threshold switching portion and a phase change memory portion with specific materials like silver-doped silicon dioxide and GST or AIST materials, respectively, to enable efficient energy use and parallel operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If a Von Neumann structure is used with separate processing unit and memory, then data processing can be performed, but bottlenecks and high energy consumption occur due to data exchange between processing unit and memory
Solution Approach 1:
The patent combines processing and memory functions into a single neuromorphic memory element that can simultaneously perform computation and data storage. The device integrates threshold switching portion and phase change memory portion in one structure, eliminating the need for separate processing units and memory, thus resolving the bottleneck of data exchange between processing unit and memory while reducing energy consumption.
Solution Approach 2:
The neuromorphic memory element performs multiple functions including data storage, data processing, and neuromorphic computation within a single device. It can emulate both neuronal plasticity (volatile storage) and synaptic plasticity (non-volatile storage), making it a universal component that replaces traditional separate processing and memory units.
2Speed
If volatile memory is used for fast data access, then processing speed is improved, but data loss occurs when power is removed
Solution Approach 1:
The patent uses a composite structure combining threshold switching material (for volatile functionality) and phase change memory material (for non-volatile functionality) in a single device. This composite approach allows the device to exhibit both volatile characteristics (fast access, threshold switching) and non-volatile characteristics (data retention without power), resolving the contradiction between speed and reliability.
3Reliability
If non-volatile memory is used for data retention, then data persistence is improved, but access speed and energy efficiency deteriorate
Solution Approach 1:
The patent utilizes phase change material that can switch between crystalline and amorphous states with different resistance values. By applying controlled voltage pulses, the device can rapidly transition between states, achieving both fast access speed and non-volatile data retention. The threshold switching mechanism further enhances speed by enabling rapid state transitions.
4Ease of manufacture
If traditional memory structures are used, then manufacturing is straightforward, but they cannot simultaneously implement volatile and non-volatile features for neuron and synapse emulation
Solution Approach 1:
The patent divides the memory element into distinct functional portions: a threshold switching portion and a phase change memory portion. Each portion has a specific function (threshold switching for neuronal plasticity, phase change for synaptic plasticity), making the device manufacturable with existing processes while achieving the versatility needed for neuromorphic computing applications.
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
This solution allows for accurate and fast learning and inference with reduced energy consumption by effectively mimicking neuronal and synaptic plasticity, achieving efficient neuromorphic computing through the neuromorphic memory element's ability to switch between resistance states based on voltage and phase changes.
Implementation Method 1
the first thin film layer may be configured to form a filament based on a magnitude of the voltage difference applied between the first electrode and the second electrode
Implementation Method 2
the second thin film layer may be configured to undergo a phase change based on a voltage pulse applied between the first electrode and the second electrode
Implementation Method 3
the second thin film layer may be configured to change phase to a crystal state when a setting signal having a first magnitude and a first width is applied between the first electrode and the second electrode, and may be configured to change phase to an amorphous state when a reset signal having a second magnitude greater than the first magnitude and a second width less than the first width is applied
Data Source
AI summary
Disclosed is a neuromorphic memory element, which includes a first electrode; a second electrode; a first thin film layer adjacent to the first electrode between the first electrode and the second electrode and that is configured to emulate a neuronal plasticity by performing a volatile storage function based on a voltage difference between the first electrode and the second electrode; and a second thin film layer between the first thin film layer and the second electrode and that is configured to emulate a synaptic plasticity by performing a non-volatile storage function.