17984750. COMPENSATION FOR CONDUCTANCE DRIFT IN ANALOG MEMORY simplified abstract (INTERNATIONAL BUSINESS MACHINES CORPORATION)
Contents
- 1 COMPENSATION FOR CONDUCTANCE DRIFT IN ANALOG MEMORY
- 1.1 Organization Name
- 1.2 Inventor(s)
- 1.3 COMPENSATION FOR CONDUCTANCE DRIFT IN ANALOG MEMORY - A simplified explanation of the abstract
- 1.4 Simplified Explanation
- 1.5 Potential Applications
- 1.6 Problems Solved
- 1.7 Benefits
- 1.8 Potential Commercial Applications
- 1.9 Possible Prior Art
- 1.10 Original Abstract Submitted
COMPENSATION FOR CONDUCTANCE DRIFT IN ANALOG MEMORY
Organization Name
INTERNATIONAL BUSINESS MACHINES CORPORATION
Inventor(s)
Charles Mackin of San Jose CA (US)
COMPENSATION FOR CONDUCTANCE DRIFT IN ANALOG MEMORY - A simplified explanation of the abstract
This abstract first appeared for US patent application 17984750 titled 'COMPENSATION FOR CONDUCTANCE DRIFT IN ANALOG MEMORY
Simplified Explanation
In an analog memory-based artificial neural network, a system is proposed to compensate for activation drift by mapping output activation vectors at different points in time. Here are the key points of the innovation:
- Input activation vectors are fed into a crossbar array at two different time points.
- Output activation vectors are read from the crossbar array at each time point.
- A function is determined to map the output activation vectors from the second time point to those from the first time point.
- This function is then applied to subsequent output activation vectors to correct for activation drift.
Potential Applications
This technology could be applied in various fields such as:
- Pattern recognition
- Signal processing
- Machine learning
Problems Solved
The system addresses the issue of activation drift in analog memory-based artificial neural networks, ensuring the accuracy and reliability of the network over time.
Benefits
The benefits of this technology include:
- Improved performance of artificial neural networks
- Enhanced stability and consistency of network outputs
- Reduction of errors caused by activation drift
Potential Commercial Applications
Potential commercial applications of this technology could include:
- Developing more efficient and reliable AI systems
- Enhancing the performance of neural network-based applications in various industries
Possible Prior Art
One possible prior art in this field is the use of calibration techniques to compensate for drift in analog memory-based neural networks.
Unanswered Questions
How does this system compare to other methods of compensating for activation drift in artificial neural networks?
This article does not provide a comparison with other existing methods for addressing activation drift.
What are the computational requirements of implementing this system in practical applications?
The article does not delve into the computational resources needed to implement this system in real-world scenarios.
Original Abstract Submitted
A system can compensate for activation drift in analog memory-based artificial neural networks. A set of input activation vectors can be input, at a first point in time, to a crossbar array. The first set of output activation vectors can be read from the output lines of the crossbar array. At a second point in time, which is a later time than the first point in time, the input set of activation vectors can be input to the crossbar array. A second set of output activation vectors can be read from the crossbar array. A function that maps the second set of output activation vectors to the first set of output activation vectors can be determined. The function can be applied to subsequent output activation vectors output by the crossbar array. A method thereof, can also be provided.
