Learn about the evolutionary algorithms that have performed well compared to other techniques in artificial intelligence.
Here's what I've written (so far):
Explore the Skip-gram model for training word vectors and learn about how negative sampling is used for this purpose.
Learn about the difference between using a hard margin and a soft margin in SVM.
Learn the difference between Instance and Batch normalization
Understand how the big-O and little-o notations differ and what it means to be asymptotically tight.
Learn about heuristic functions, their benefits and pitfalls, and some of the examples where we can use them.