Analytics Summit Speakers & Bio's

David Hand.png

Plenary talk: Prof David Hand presents AI validation, reliability and maintenance

Professor David Hand will present a high-level view of requirements for valid, reliable, and robust AI systems, starting with the purpose of the system. He will examine the risks arising from inadequate data and outline how to overcome them, and discuss challenges that arise from the fundamental nonstationarity of the world, the need for AI systems to work in a human social context, and the need for them to work in an AI context, especially as the internet of things becomes more dominant. Finally, Professor Hand will look at the role of explainability in trustworthy systems, and the use of checklists to create them.

About Professor David Hand
Professor David Hand is a co-proposer of the Validate AI Conference. He is also Senior Research Investigator and Emeritus Professor of Mathematics at Imperial College, London, where he previously held the Chair of Statistics. He serves on the Board of the UK Statistics Authority and the European Statistical Advisory Committee and is a former president of the Royal Statistical Society. Professor Hand has received many awards for his research and has published 29 books.

David FosterMark Somers presents Fair Decisions

This talk examines why it is hard to build models, particularly using AI and Machine Learning techniques, that deliver fair decisions. Much emphasis has been placed on being able to interpret models so that we can ensure they meet the model owners standards of fair treatment and that management can be held responsible for their outcomes. However this is a limited approach – either the model is limited to what humans can understand, reducing any benefit in applying more sophisticated methods, or the interpretability is an illusory decoration and the underlying model may not necessarily be fair at all. This talk begins to explore some ideas of how to build models that are fair by design – this requires us first to agree on what we mean by fairness and then decide how best to reflect this in a model architecture. We will consider how the assessment of fairness can be re-configured into a set of conditions that can be largely automated and potentially embedded in the model training process itself. Ultimately fairness is a subjective assessment and societies subjective views will change over time, however as a model owner responsible for the outcomes of automated decisions the best rebuff to challenges to your assessment is to be transparent and precise about what you mean by fairness and how you have implemented it within the decision process.

About Mark Somers
Mark Somers is Managing Director at 4most, where his role is to inspire creativity, encourage cohesive teamwork and embed an analytical culture at the core of the company. He has a track record of building successful analytical teams. Mark has an in-depth knowledge of the credit risk regulatory landscape and statistical and machine learning techniques, and he has practical experience of implementing complex regulatory change in technical areas, such as Basel and IFRS9, balancing both demonstrable good governance and the need for business agility.

David Foster

Ganna Pogrebna presents The Data Science of Hollywood: From Content Generation to AI-driven Supply Chains

Improving productivity and customer experience in the entertainment industry is challenging, in part because the industry relies heavily on a consumer-centric framework that puts customers at the centre of content development and production.

This talk will propose a new framework which allows to use data science to optimise content-generation in entertainment. Pogrebna will use natural language processing, image recognition as well as advanced sound analytics methodology, combined with econometric analysis, to explore whether and to what extent human emotions, imagery format, and music patterns shape consumer preferences for media and entertainment content, which, in turn, affect revenue streams.

About Ganna Pogrebna
Ganna Progrebna is a Professor of Behavioural Economics and Data Science at the University of Birmingham and a Fellow at the Alan Turing Institute. She is also a Turing Lead for the University of Birmingham and a Behavioural Data Science Lead at the Alan Turing Institute. Blending behavioural science, computer science, data analytics, engineering, and business model innovation, Ganna helps businesses, charities, cities, and individuals to better understand why they make decisions they make and how they can optimise their behaviour to achieve higher profit, better cybersecurity, more desirable social outcomes, as well as flourish and bolster their well-being.

David Foster

Jack Snape presents Making Sense of Simulated Worlds: Using the Principles of Predictive Analytics to Make Complex Simulation Models More Usable

Simulation models are moving closer to 'real world' levels of detail and being linked together into complex ecosystems, due to need to understand interactions between sectors, such as transport, land-use and electricity generation.

In this talk, Jack Snape will share his experience using principles of predictive analytics to make complex modelling ecosystems more usable. He will also outline the principles of the approach and set out a current example - a simplified representation of the North’s road and rail network, carefully calibrated to detailed transport models - and provide an overview of future developments, including the potential to use in freight modelling and with agent-based models.

About Jack Snape
Jack has a PhD in physics and has worked as an analyst in the Civil Service and in Local Government across a range of policy areas, including higher education, manufacturing, climate change and transport. At the Committee on Climate Change he was responsible for advising the Government on reducing carbon emissions from transport in the UK. Jack is now the Analysis Manager at TfN, leading the development of TfN’s Analytical Framework, a new suite of software tools that provides a consistent approach to data, modelling and appraisal across travel modes and regions of the North.

Paul LaughlinPaul Laughlin presents Developing the Softer Skills Analysts need to make a difference in your business

In this talk, Paul will present an introduction to the need for softer skills development for analysts and data scientists. He will walk us through his 9-step Softer Skills model and provide advice for leaders in developing the capabilities of their teams, including guidance on bespoke competency frameworks for analysts and developing the culture needed to support effective agile working. All talk attendees will receive a digital copy of the 9-step Softer Skills model and action commitments to use in their workplaces.

Paul Laughlin
With over 20 years of experience in creating and leading data & analytics teams, Paul now focusses on helping other leaders master "the people side of data and analytics" with his consultancy, Laughlin Consultancy.

Paul has published both academic research and opinion from his commercial experience. These include chapters in books such as 'The Dark Side of CRM' and numerous articles in journals including 'The Bottom Line'. 

Paul also collaborates with a number of UK universities, chiefly the University of South Wales. He is on the panel of coaches in their Welsh Coaching Centre and guest lectures on MSc programmes, including MSc Data Science & MSc Financial Services Management.

David Foster

Marilena Karanika presents categorisation in the age of Open Banking

This talk will discuss the Open Banking initiative and the opportunity it presents: i) for consumers to use their transactional data to access better financial products and services and ii) for businesses to improve their decision making processes and provide their customers with the best possible experience.

Deploying a transaction categorisation engine is crucial for transforming data into insights and obtaining a holistic customer view. Specific use cases across different points in the customer journey will be covered to illustrate the power of this data source alongside other more traditional data feeds.

About Marilena Karanika
Marilena Karanika is Head of Categorisation at Experian, providing data and analytics support across different product verticals. With more than 10 years of experience in credit risk modelling and analytics, a key area of expertise is enabling financial organisations make better use of their data to reach more informed decisions and appropriately support their customers throughout their customer journeys.

Marilena is passionate about linking academia and industry and works closely with a number of universities across the UK for career days, guest lectures in MSc programs in Data Science, Banking & Risk, and Business Analytics, as well as MSc summer placements.