Blog

I used to use Datomic, https://www.datomic.com, a lot. Datomic is a unique database. The most attractive feature would be the idea of preserving transactions. Based on the transaction id, Datomic is able to pull out the past data which has been already updated later.

This post is going to be very different from what I write here. The content is a memo how I create a GCP (Google Cloud Platform) instance for Deep Learning. While I study algorithm stuff, I also have been studying Deep Learning. In my early days, I tried to train my Deep Learning model only on a laptop. My laptop is 2012 model MacBook Pro, so I would say it is reasonably fast. However, when it comes to Deep Learning, the training was quite painful on the such machine. Often, I ran the training over night to get a disappointed result. Still, I didn’t use any of pricey cloud environment since it was my personal study unrelated to my day job. I wanted to save money.

Problems which ask a binary tree traverse, add/delete nodes, etc. are popular in algorithm questions. The binary trees are often just a binary tree or binary search tree. Sometimes, the problem pinpoints a particular type of a binary tree, for example, a balanced binary tree or complete binary tree.

What is Catalan number ? The Catalan number belongs to the domain of combinatorial mathematics. It is a sequence of natural numbers such that: 1, 1, 2, 5, 14, 42, 132, 429, 1430, 4862, 16796, 58786, 208012, 742900, 2674440, 9694845, 35357670, 129644790, 477638700, 1767263190, ... The sequence appears in counting problems. Wikipedia has the details about the sequence: Catalan Number.

“Given an integer n, find whether n is a prime number or not” – This is a typical algorithm question. If the given number is small, repeating a division from 2 to n gives the answer.