Artifcial Intelligence Data Science Technology

This is the first in a regular series we’ll be running from here on in, looking at some of the interesting and diverse background of people working at Unico.  Today let’s look at the role of a Data Scientist so we’re talking to Ross Ashman PhD our Head of Data Science at Unico on why his first AI job examining the authenticity of paper made him the ideal dinner party guest in the mid 2000’s.

What is a Data Scientist?

 

So, Ross what does a Data Scientist do?

Essentially Data Scientists use scientific method to help business solve complex problems or find the answer to complex questions by knowing how to unlock, analyse and model datasets.

Google coined the term data scientist, driven by them and how they analyse all the information they have to target ads, searches etc.

What I do as Lead Data Scientist at Unico is build Artificial intelligence models from the ground up that solve business problems for our customers, so for example we’re currently working on a  signature verification tool that can automate signature verification and do it faster than a human, to teach a machine to do that is not a straight forward task but data scientists love working on complex problems and trying to solve them.

 

What was your pathway into Data Science?

I studied Physics and Chemistry at Flinders University and was drawn to working with lasers, so I did an Honors project using lasers to look at chemical interactions. After that I did a PhD at the University of Western Australia in Perth.

My PhD was in Biomedical Optics, using lasers to image and investigate the human eye. I continued this into my post doc at Melbourne University where I did research on adaptive optics for high resolution retinal imaging and psychometric testing. The primary aim being to stimulate single rods and cones in the eye to understand the visual paths ways from the retina to the brain.

I stayed in academia for a while but soon realised it wasn’t for me, the hand to mouth existence of applying for limited grants and funding didn’t feel like it was getting me anywhere.  So, I saw a job in advertised at Cannon in image processing and applied, thus beginning my corporate career.

I realised as soon as I started, this was where I belonged, instead of being rudderless and left to your own devices, there was structure, the role was research focused with a goal a plan and an endpoint in mind.

My first exposure to Machine Learning was in a document security project, where we used machine leaning to identify paper types from scanned images of documents. I produced an experiment and the analysis to demonstrate that a technique developed at Canon, called paper finger printing, could be used to uniquely identify single sheets of paper from scanned images.  The purpose was to establish a ‘paper fingerprint’ and identify whether a print was the original or a reproduced scan.  This was the start of my Machine Learning and AI career.

 

What are some of your career highlights so far?

I’ve been lucky enough to work in a variety of roles and organisation in my Data Science career but there are several highlights.

Firstly, in the DSTO (now DSG) looking at advanced technologies and what the implications were for defence was very interesting. In my role there I researched and examined new technology such as quantum computing, nano-technology and 3D printing to assess and present the defence potential threats we needed to be aware of.  I built a system that would trawl the literature and papers on specific technologies and combine with briefings we would have with expert professors in the field of emerging technology.

Secondly at EY I was part of a team that built an application platform for retail data and analytics recommendations, the application allows customer segmentation and analysis together with a single customer view recommender system. It also included a price elasticity model. These analyses and predictions can be integrated into target lists for marketing campaigns and interventions to prevent customer churn.

However I’m excited to say I’m doing some of my most interesting work to date right now at Unico, using AI to solve some of the challenges in business such as the manual work required for a human to eyeball a signature on a paper form.  The projects combine my two favorite areas; imaging and text. The fact that technology can solve some of these problems and learn how to do it faster and more accurately is really exciting to me.

 

What is your dream client brief?

Going back to my background, it’s all around combining image and language processing to solve common problems.  For me, the more challenging the problem and the more benefit for the client, the better.

 

How do you think Data Science will change the world?

Notwithstanding security and privacy concerns I think there is enormous capacity  for data science and artificial intelligence to improve every aspect of our lives from Healthcare to customer experience and everything in between but I believe as is the way of most technologies, there will be a societal input as to how these are adopted or rejected as part of our co-existence with technology and the part we want it to play in our lives.