Backgound in Electromagnetic Theory, Plasma Physics and Radar Target Identification using Neural Networks.

PhD Student, Research Assistant and Lecturer at Stanford;

AB, ScB, ScM Student; Research Assistant, Fellow and Professor at Brown;

27 yrs researching Ballistic and Theatre Missile Defense using Neural Networks at MIT Lincoln Laboratory. Retired 2003.

PLEASE DO NOT SEND QUESTIONS AND DATA TO MY EMAIL. HOWEVER, CAN SEND LINKS TO POSTS.

Professional Interests: Neural Netwoks, Spectral Analysis

Answered

Would this be considered underfitting?

A model is UNDERFIT if and only if No. of independent training equations < No. of unknowns Hope this helps Tha...

Would this be considered underfitting?

A model is UNDERFIT if and only if No. of independent training equations < No. of unknowns Hope this helps Tha...

4 meses ago | 0

Answered

Avoid exploding/vanishing gradient problem with NARX nets?

Use a higher open-loop peformance goal. Then lower the value after the loop is closed. It's been years since I've done this so ...

Avoid exploding/vanishing gradient problem with NARX nets?

Use a higher open-loop peformance goal. Then lower the value after the loop is closed. It's been years since I've done this so ...

5 meses ago | 0

Answered

How to make prediction from a trained NARX neural network?

You forgot to include the intial conditions: yz = nets(xz,xiz,aiz); Thank you for formally accepting my answer Greg

How to make prediction from a trained NARX neural network?

You forgot to include the intial conditions: yz = nets(xz,xiz,aiz); Thank you for formally accepting my answer Greg

5 meses ago | 0

Answered

Out of memory issue while training a Neural Network (NN), array exceeds maximum array size preference using backpropJacobianStatic

A single hidden layer is sufficient. Hope this helps Thank you for formally accepting my answer Greg

Out of memory issue while training a Neural Network (NN), array exceeds maximum array size preference using backpropJacobianStatic

A single hidden layer is sufficient. Hope this helps Thank you for formally accepting my answer Greg

6 meses ago | 0

Out of memory issue while training a Neural Network (NN), array exceeds maximum array size preference using backpropJacobianStatic

A single hidden layer is sufficient. Hope this helps Thank you for formally accepting my answer Greg

6 meses ago | 0

Answered

Difference between function of sim and predict in Neural network?

help sim doc sim help predict doc predict Thank you for formally accepting my answer Greg

Difference between function of sim and predict in Neural network?

help sim doc sim help predict doc predict Thank you for formally accepting my answer Greg

6 meses ago | 0

Answered

Forecasting using GRNN (Generalized Regression Neural Network)?

Yes, Normalize input and output data to [-1,1]. Minimize the order for a sufficient polynomial fit. Minimize the number of ...

Forecasting using GRNN (Generalized Regression Neural Network)?

Yes, Normalize input and output data to [-1,1]. Minimize the order for a sufficient polynomial fit. Minimize the number of ...

6 meses ago | 0

Answered

CNN With unbalanced Data

Although you do not prefer data augmentation, duplication of the smaller dataset examples is probably the quickest and most reli...

CNN With unbalanced Data

Although you do not prefer data augmentation, duplication of the smaller dataset examples is probably the quickest and most reli...

7 meses ago | 0

Question

How to find questions on neural neural networks.

I have been away for awhile: How can I list the neural network questions? Greg

7 meses ago | 1 answer | 0

Answered

Trained neural network to optimize input variable

"OPTIMIZED INPUT" is ill-defined Just train another net x2 = f2(x1,y1,y2) Hope this helps, THANK YOU FOR FORMALLY ACCEP...

Trained neural network to optimize input variable

"OPTIMIZED INPUT" is ill-defined Just train another net x2 = f2(x1,y1,y2) Hope this helps, THANK YOU FOR FORMALLY ACCEP...

8 meses ago | 0

Answered

Linear Integer Output from a Neural Network

The integer nature of the output DOES NOT HAVE TO BE CONSIDERED during training. It is sufficient to round the real valued ou...

Linear Integer Output from a Neural Network

The integer nature of the output DOES NOT HAVE TO BE CONSIDERED during training. It is sufficient to round the real valued ou...

8 meses ago | 0

Answered

How to disable validation and test data set in neural network

You have to define net before modifying any properties. clear all, close all, clc [x,t] = iris_dataset; for i = 1:2 net ...

How to disable validation and test data set in neural network

You have to define net before modifying any properties. clear all, close all, clc [x,t] = iris_dataset; for i = 1:2 net ...

11 meses ago | 1

Answered

Training data and Training target in Neural Networks

You cannot make any intelligent decisions until you have examined a plot of the data!!! (WRONG!!! Plotting the data first is ...

Training data and Training target in Neural Networks

You cannot make any intelligent decisions until you have examined a plot of the data!!! (WRONG!!! Plotting the data first is ...

11 meses ago | 0

| accepted

Answered

simulink neural network producing different outputs to workspace

A simpler solution is to ALWAYS begin the program with a resetting of the random number generator. For example, choose your favo...

simulink neural network producing different outputs to workspace

A simpler solution is to ALWAYS begin the program with a resetting of the random number generator. For example, choose your favo...

12 meses ago | 1

Answered

How to avoid getting negative values when training a neural network?

Use a sigmoid for the output layer. Hope this helps THANK YOU FOR FORMALLY ACCEPTING MY ANSWER GREG

How to avoid getting negative values when training a neural network?

Use a sigmoid for the output layer. Hope this helps THANK YOU FOR FORMALLY ACCEPTING MY ANSWER GREG

12 meses ago | 0

Answered

weird plotregression plots for a 10% of my fitnet neural networks

Sometimes training gets into a parameter space rut. That is why it is wise to train multiple models. Hope this helps. Greg ...

weird plotregression plots for a 10% of my fitnet neural networks

Sometimes training gets into a parameter space rut. That is why it is wise to train multiple models. Hope this helps. Greg ...

12 meses ago | 0

Answered

What is the purpose of shuffling the validation set?

To impose and verify a consistent GENERALIZED path to convergence by avoiding repetitive anomalies. Hope this helps Greg

What is the purpose of shuffling the validation set?

To impose and verify a consistent GENERALIZED path to convergence by avoiding repetitive anomalies. Hope this helps Greg

12 meses ago | 0

Answered

How to force overfiting of Deep Learning Network for Classification

OVERFITTING = More training unknowns (e.g., weights) than training vectors. OVERTRAINING1 = Training an overfit network to...

How to force overfiting of Deep Learning Network for Classification

OVERFITTING = More training unknowns (e.g., weights) than training vectors. OVERTRAINING1 = Training an overfit network to...

12 meses ago | 0

Answered

Neural Network Pattern Recognition

Targets are 1-dimensional unit vectors with 4 zeros and a single 1 . Thank you for formally accepting my answer. Greg

Neural Network Pattern Recognition

Targets are 1-dimensional unit vectors with 4 zeros and a single 1 . Thank you for formally accepting my answer. Greg

12 meses ago | 0

| accepted

Answered

semanticseg producing marginally different values when inference is repeated

Use clear all, close all, clc, rng(0) on the 1st line Thank you for formally accepting my answer Greg

semanticseg producing marginally different values when inference is repeated

Use clear all, close all, clc, rng(0) on the 1st line Thank you for formally accepting my answer Greg

alrededor de 1 año ago | 0

Answered

combining two neural networks (net1 is trained & net2 is untrained) in one bigger network

Just 1. Save the outputs of net1 in a file 2. Use the file to train net2 Greg

combining two neural networks (net1 is trained & net2 is untrained) in one bigger network

Just 1. Save the outputs of net1 in a file 2. Use the file to train net2 Greg

alrededor de 1 año ago | 0

Answered

Problem with the TreeBagger Command

The sizes of the input function and output target must be [ I N ] = size(input) [ O N ] = size (target) Hope this helps...

Problem with the TreeBagger Command

The sizes of the input function and output target must be [ I N ] = size(input) [ O N ] = size (target) Hope this helps...

alrededor de 1 año ago | 0

Answered

How do I create a neural network that will give multiple input and outputs?

ALWAYS arrange your data so that [ I N ] = size(input) [ O N ] = size(output) Hope this helps. Greg

How do I create a neural network that will give multiple input and outputs?

ALWAYS arrange your data so that [ I N ] = size(input) [ O N ] = size(output) Hope this helps. Greg

alrededor de 1 año ago | 0

Answered

How to make a hybrid model (LSTM and Ensemble) in MATLAB

Replace your YES/NO data with either 1/0 or 1/-1. Hope this helps. Greg

How to make a hybrid model (LSTM and Ensemble) in MATLAB

Replace your YES/NO data with either 1/0 or 1/-1. Hope this helps. Greg

alrededor de 1 año ago | 1

Answered

how to augment image data only for a specific class?

Separate class 0 and interpolate. If you have a good feel for the data you could extrapolate. However the latter might be tricky...

how to augment image data only for a specific class?

Separate class 0 and interpolate. If you have a good feel for the data you could extrapolate. However the latter might be tricky...

alrededor de 1 año ago | 0

Answered

NTSTOOL - How to get predicted values of "the future"?

The known time series is analyzed to yield a time-series model that uses past and present values to predict future values. The ...

NTSTOOL - How to get predicted values of "the future"?

The known time series is analyzed to yield a time-series model that uses past and present values to predict future values. The ...

más de 1 año ago | 0

Answered

Timedelaynet output calculation principle

You did not include tHe 2 biases. Hope this helps. Greg THANK YOU FOR FORMALLY ACCEPTING MY ANSWER

Timedelaynet output calculation principle

You did not include tHe 2 biases. Hope this helps. Greg THANK YOU FOR FORMALLY ACCEPTING MY ANSWER

más de 1 año ago | 0

Answered

Understand number of weights of Neural Network

It is possible. In general, however, you don't have the slightest idea what choice would be significantly better than random. ...

Understand number of weights of Neural Network

It is possible. In general, however, you don't have the slightest idea what choice would be significantly better than random. ...

más de 1 año ago | 0

Answered

Using pca for features selections

PCA (Principal Coordinate Analysis) is a very useful method for regression (it ranks linear combinations of the original variabl...

Using pca for features selections

PCA (Principal Coordinate Analysis) is a very useful method for regression (it ranks linear combinations of the original variabl...

más de 1 año ago | 0

Answered

How do we decide the number of hiddenlayers in a PatternNet?

patternnet(10) indicates ONE HIDDEN LAYER WITH TEN NODES It is important to be mindful of the number of layers and nodes. The ...

How do we decide the number of hiddenlayers in a PatternNet?

patternnet(10) indicates ONE HIDDEN LAYER WITH TEN NODES It is important to be mindful of the number of layers and nodes. The ...

más de 1 año ago | 0

| accepted