Machine Learning

Machine Learning is not a new subject but is becoming an increasingly hotter topic. We have three speakers to each bring a different spin on the subject and therefore it should be a night where there will be something for everyone, whether you are Data Scientist, Analyst or Digital Marketer.

In no particular running order

Harvinder Atwal - Money Supermarket

Harvinder leads a team of analysts to deliver data-driven customer insight, marketing optimisation and predictive analytics for Moneysupermarket from one of the largest customer databases in the UK with records for 23 million unique individuals.

Harvinder was previously Insight Director for Tesco Clubcard at Dunnhumby and Senior Manager for Customer Strategy and Insights at Lloyds Banking Group.He is passionately interested in Data Science, Machine Learning, Big Data technologies and how they can be used to improve Customer Experience.

Harvinder’s presentation is called Machine Learning – What they don’t teach you on Coursera

Harvinder will be showing some live models used at MoneySuperMarket. However, delving into how to successfully implement Machine Learning rather than just having great algorithms. Including, selling the benefits to the business, working with developers and engineers to put the model into production, building testing into the process and ongoing maintenance of the solution.

Nick Walker will be taking a different slant and be talking more about deep learning. Which will include a live demo. Having seen Nick do a Live demo before at MeasureCamp London I am quite excited about this.

Nick Walker

Nick was introduced to machine learning at university, then while working in aerospace and banking in the mid 90s he got his first chance to experiment with neural networks. And he’s had an obsession with them ever since. Nick is now a freelance consultant specialising in bespoke development, integration and analytics for commerce and more recently using machine learning techniques in analytics and healthcare applications.

Deep Learning with Nick

He will be talking about Deep Learning, the area of machine learning that broke out of the research labs in the last couple of years with real life applications such as computer vision and translation. He will talk about current uses of Deep Learning, a live demonstration and talk about potential future applications in analytics.

Richard Fergie- Consultant at E-Analytica

Richard has worked in digital since 2008. After working for various agencies in SEO. Richard is now a consultant as is in constant demand for his consultancy work in PPC and Analytics.

Hierarchical models

One of the challenges of paid search is capturing everything we know and all our assumptions about the market when setting a bid. Hierarchical models is one way of approaching this and you don't have to get very far into this before markov chain monte carlo methods start to become important.

This methodology gives a distribution estimate rather than a point estimate for model parameters. This has wider implications in web analytics as we try to make better decisions - a bet with a 1/100 chance of winning will lose on average, but it is still worth taking if the payoff is big enough.