Recently I have been working on the Synaptic Transistor for Neuromorphic Application; for a more detailed understanding of the Neuromorphic Application, I would like to learn to simulate an artificial neural network (ANN) to perform supervision on the Modified National Institute of Standard and Technology (MNIST) handwritten database. Furthermore, I would like to calculate the pattern recognition accuracy of the prepared for following synaptic transistor parameters. Therefore, can you kindly suggest how to write a Tensorflow python program? Thanks!
Key parameter:
Multi-layer perceptron (MLP) network:
784(input neurons),
256 (Hidden layer),
10(output neurons),
Our device exhibited a potentiation/depression curve with a high Gmax/Gmin value (∼14.1), conductance states with 68 levels, and suitable linearity (Ap, 0.85; Ad, 0.98).
Aparna Sathya Murthy, Thanks for your response; I have some experimental results, such as a synaptic weight potentiation/depression curve with a high Gmax/Gmin value, conductance states levels, and suitable linearity; I would like to use these values in the TF variable in the MINIST database to calculate handwritten image accuracy recognition. However, I am not familiar with using coding! Can you suggest using the variable in TensorFlow to calculate image accuracy recognition? For your information, I have attached the MINIST database for image recognition. Thanks, Regards
Aparna Sathya Murthy, Thank you for your information and help; I also face the same problem! My work is almost similar to the above you referred paper.
Aparna Sathya Murthy, Apology, I am not sure exactly what you mean. However, I need to simulate the ANN handwritten recognition accuracy for my experimental work, such as synaptic weight and Non-linearity values. So I need to write coding for these values into the typical ANN recognition accuracy coding to calculate the accuracy recognition.
- Badges
- Science topic
- Similar topics
- Chemistry
- Nanotechnology
- Nanostructures
- Thin Films