Exam DP-100 topic 5 question 19 discussion

I agree, B and C should be the correct answers

adding more training records should decrease the overfitting.

Explanation: B. L1/L2 regularization helps prevent overfitting by adding a penalty term to the loss function, discouraging the model from relying too heavily on any particular feature. C. Data augmentation increases the diversity of your training set by applying random (but realistic) transformations to the existing images, which helps the model generalize better and reduce overfitting. These two techniques are commonly used to address overfitting in deep learning models, especially CNNs for image classification.

B. Add L1/L2 regularization. C. Use training data augmentation. These methods directly address the problem of overfitting by either penalizing overly complex models or by making the training data more diverse and challenging for the model.

This reference might be useful: https://towardsdatascience.com/8-simple-techniques-to-prevent-overfitting-4d443da2ef7d

The two actions that can help reduce overfitting and converge the model to an optimal fit are: B. Add L1/L2 regularization: Regularization techniques can help to reduce overfitting in a neural network. L1/L2 regularization adds a penalty term to the loss function, which encourages the model to learn simpler and smoother weight values. This, in turn, helps to prevent overfitting. C. Use training data augmentation: Data augmentation is a technique that can be used to artificially increase the size of the training dataset by creating new examples from existing data. This can help the model to generalize better and reduce overfitting. Common data augmentation techniques for image data include random rotations, flips, and translations. Options A and E suggest adding additional dense layers, which can increase the complexity of the model and potentially exacerbate overfitting. Option D suggests reducing the amount of training data, which can lead to underfitting and poor generalization performance. Therefore, options B and C are the best choices for reducing overfitting and improving model performance.

Regulation Increase data through data argumentation

I agree with B and C

Answer should be B and C

Regularisation and data augmentation are correct. Dropouts and early termination are also correct but not in the options.

During dropout, we are not actually reducing the training data but rather dropping the neurons to help them memorize less and not overfit. https://www.kdnuggets.com/2019/12/5-techniques-prevent-overfitting-neural-networks.html so the answer is B and C

Answer should be B,C

Yes BC should be correct, image data augmentation - generating more training data artificially to expand the learning of algo

and to support the argument for C: "Image data augmentation is a technique that can be used to artificially expand the size of a training dataset by creating modified versions of images in the dataset. Training deep learning neural network models on more data can result in more skillful models, and the augmentation techniques can create variations of the images that can improve the ability of the fit models to generalize what they have learned to new images." Ref: https://machinelearningmastery.com/how-to-configure-image-data-augmentation-when-training-deep-learning-neural-networks/

Yes, BC are correct.

BC are right answers. To reduce overfitting in DL model, you either increase training data volume or reduce complexity of the model

The answer is definitely B and C