Clik here to view.
Explain to Me: Why Train A Model Generatively and Discriminatively
Summary These are my thoughts on why we want to train a model generatively and discriminatively (or why we want to train a model using unsupervised fashion and then supervised fashion): We want to make...
View ArticleComputational Challenges in Precision Medicine and Genomics
A great slidedeck describing just in the right detail for those who’s interested in applying machine learning and computational disciplines to the field of biology, medicine, and genomics. It talks...
View ArticleDo You Re-train on the Whole Dataset After Validating the Model?
Suppose we have a dataset split into 80% for training and 20% for validation, do you do A) or B? Method A) Train on 80% Validate on 20% Model is good, train on 100%. Predict test set. Method B) Train...
View ArticleClik here to view.
How to Select Your Final Models in a Kaggle Competition
Did your rank just drop sharp in the private leaderboard in a Kaggle competition? I’ve been through that, too. We all learn about overfitting when we started machine learning, but Kaggle makes you...
View ArticleClik here to view.
Visualizing the Differences In L1-norm and L2-norm Loss Function
In an earlier post about the differences between L1 and L2 as loss function and regularization, one of the graph about L1-norm and L2-norm loss function is rather confusing to many readers, as I have...
View ArticleClik here to view.
Why is Keras Running So Slow?
More notes for myself… so it may not be helpful for you who bumped into here. Why This Article? Setting Theano correctly is not enough to ensure you can run deep learning software correctly. In our...
View ArticleClik here to view.
How Does the Number of Hidden Neurons Affect a Neural Network’s Performance
.highlight{background: #f8f8f8; overflow:auto;width:auto;border:solid gray;border-width:.1em .1em .1em .1em;padding:0em .5em;border-radius: 4px;} .k{color: #338822; font-weight: bold;} .kn{color:...
View Article
