Machine Learning

on November 1, 2018

I’ve been working in the data and analytics space since I was a fresh out of university. I got a job as a data analyst at Croyden Council in London working with huge volumes of data and Microsoft Excel. I was lucky to start my job at Croyden Council a few months before a 4 yearly report was due. Their data collection and analysis systems were not as automated as they would have liked and I got the opportunity to create a large amount of automated analysis and collation software while I was there. I’ve been enamored with the value data provides since then.

Whenever I’ve built or worked in an organisation since, extracting, collating and analysing data has been an important part of the organisations success. The insights that data can provide you and the organisation are invaluable.

I stumbled across Machine Learning a while ago and was in awe of its power both now and the possibilities it holds as the technology and mathematics progresses. The beauty of the mathematics that underlies it, is like fine art, it is beautiful in both its simplicity and complexity.

The initial learning curve was no different to anything else I have ever encountered. It seemed complicated to start with but once I got my head around the math, things started to fall into place. I had been very lucky to have spent the previous 2 years studying human behavioral psychology and brain science. These lessons helped me to understand the underlying logic in regards to why the math was structured the way it is and how it mirrors many of the functions within the human brain.

Since discovering Machine Learning I have worked on a number of projects that use machine learning to solve both repetitive task and tasks that humans cannot cognitively process. It is obvious to most of us that machine learning and robotics are great at doing repetitive tasks or dangerous tasks, the low hanging fruit so to speak. It seems to be less obvious that there are many tasks out there that humans are simply unable to do because we do not have the cognitive power to process the information. Some of the projects I have or are currently working on which use machine learning are:

  1. Access Privilege analysis for Rulerr
  2. Loss Prevention in retail stores
  3. Asset Management
  4. Video analysis and stitching and its use in free to air and on demand broadcasting