Data Projects and Tools

A couple of members will share some of their experience with their tools or particular tools. It will run a bit like this:-

5:00 Introduction and Welcome coffee and tea in the basement.
5:20 25 minutes on a members project
5:45 break to visit bar if required
6:00 25 minutes on a tool a member uses.
6:25 Concluding discussion, date of next meeting
6:30. Retreat to 'Folk' (onsite bar) for those who wish to.

The first session with be Lead Agency's Paul Byrnes looks at machine learning in pathology slide imaging.

The second session will be Village Software's Dom Bisset having a Hello World look at the Azure Machine Learning Studio.

The session will be held in the basement meeting rooms at The Tempest, if you arrive on time come to the foyer and you will be directed down. If you arrive late enter via 'Folk' go out the side door into the Foyer and head down to the basement (we'll leave out a sign).

Session 1

My name is Paul Byrnes and I’m a data scientist in the data team at The Lead Agency. I’ve recently finished a PhD in Statistical Machine Learning at the University of Liverpool.

Visual inspection of Whole Slide Images (WSI) is commonly used for the identification of cancerous cells. During this process, a biopsy sample is acquired from a patient before being evaluated by a pathologist through the use of a microscope. Intuitively, two issues for both patient and practitioner arise during this process. Firstly, visual inspection is extremely time consuming. Given the aggressive nature of the disease, time has been identified as a key factor in improving survival rates. Theoretically if correctly implemented, machine learning has the potential to reduce the waiting time between obtaining the biopsy sample and returning the determined diagnosis. In turn, allowing for possible treatments to be started at an earlier time. The second issue with visual inspection is the possible variation in diagnosis between pathologists. This talk explores the possibility of machine learning in offering a tool for pathologists to improve the decision making process for breast cancer diagnosis.

Session 2

Dom Bisset, is Village Software's Senior BI Analyst and will be hosting the event.

Dom will demonstrate an Azure Machine Learning Studio, hello world application, predicting physical machine speeds on a production line.