Author archive for Nikolay Manchev, page 2

Data science is facing an overwhelming demand for CPU cycles as scientists try to work with datasets that are growing in complexity faster than Moore’s Law can keep up. Considering the need to iterate and retrain quickly, model complexity has been outpacing available compute resources and CPUs for several years, and the problem is growing quickly. The data science industry will need to embrace parallelization and GPU processing to efficiently utilize increasingly complex datasets.

In a previous blog post we talked about the foundations of reinforcement learning. We covered classical and operant conditioning, rewards, states, and actions, and did a review of some common reinforcement learning use-cases. This entry is a continuation of the series. In it, we present the k-armed bandit problem - a very simple setting that enables us to introduce the interaction between some of the key components of reinforcement learning.

Introduction

In this blog post we give a quick introduction to Ray. We talk about the architecture and execution model, and present some of Ray's core paradigms such as remote functions and actors.

Apache Spark, Dask, and Ray are three of the most popular frameworks for distributed computing. In this blog post we look at their history, intended use-cases, strengths and weaknesses, in an attempt to understand how to select the most appropriate one for specific data science use-cases.