Data. Uncertain.

9-9.05am

Welcome from the Centre Manager

Jessica Cornock is the Centre Manager of the ARC Industrial Transformation Training Centre – DARE. Jessica previously held the position of Operations Manager at the Centre for Translational Data Science (CTDS) and also worked at the University of Sydney as a Consultant in Research Technology, ICT. Prior to joining the University in 2013, she worked in stem cell research, pre-clinical drug development for a pharmaceutical company and clinical data management for clinical trials.

9.05-9.10am

Welcome to Country

910-9.30am

Opening remarks

Professor Philippa Pattison AO

Professor Philippa (Pip) Pattison is Deputy Vice-Chancellor (Education) at the University of Sydney and a member of the DARE Centre Advisory Board. Pip is responsible for the University’s strategy and vision for teaching and learning and students’ educational experience. Pip was named on the Queen’s Birthday 2015 Honours List as an Officer of the Order of Australia for distinguished service to higher education, particularly through contributions to the study of social network modelling, analysis and theory, and to university leadership and administration.

Professor Sally Cripps

Sally Cripps is Director of the Data Analytics for Resources and Environments (DARE) Centre. She was awarded her PhD in 2002 and her work has appeared in the world’s most prestigious statistical journals. Her research has focused on the development of new and novel methods to flexibly model and analyse complex data.

9.35-9.55am

Anecdotal thinking in health and business

When making decisions about health or finances, there is an abundance of information to draw on. Two primary kinds of information are statistical, based on large samples and that communicate uncertainty, or anecdotal, reflecting a single case. My research shows how anecdotes have large effects on our decisions despite statistical information that should render the anecdotes unconvincing. Further results show that multiple methods designed to reduce the influence of anecdotes had no such effects, and that deciders are influenced by anecdotes because it appears they think they should be influenced by them; it is not simply a cognitive failure to ignore them. After presenting the results I hope to discuss with the audience ideas for how to mitigate the effects of anecdotes.

Speaker: Dr Micah Goldwater

Dr Goldwater is a Senior Lecturer at the School of Psychology, University of Sydney. He received his BA in Linguistics from the University of Rochester in 2003, completed his PhD in Psychology at the University of Texas at Austin in 2009, and then held a Postdoctoral Fellowship at Northwestern University until joining the University of Sydney in 2013.

9.55-10.15am

Embracing variation and accepting uncertainty

It is said that your most important collaborator is yourself in 6 months.  Perhaps the best way to improve our communication of data uncertainty to others is to learn how to better communicate to ourselves.  What does it mean to say to future-you:  “I don’t know”?  Or, even more challenging, “I know a little but I’m not completely sure”?  We will discuss in the context of applications in drug testing, election forecasting, and the evaluation of scientific research.

Speaker: Professor Andrew Gelman is a Professor of Statistics and Political Science and Director of the Applied Statistics Center at Columbia University. He has received the Outstanding Statistical Application Award from the American Statistical Association, the award for best article published in the American Political Science Review, and the Council of  residents of Statistical Societies Award for outstanding contributions by a person under 40.

10.20-10.40am

Using data to help us ask the right questions

Most organisations and people working in those organisations operate in the face of uncertainty. I’ve worked in commercial organisations where the take-up of products and services are unknown, but uncertainty is mitigated through competitor analysis, product testing – beta launches, focus groups. In social policy we are faced by similar issues but I believe hugely exacerbated because the issues and drivers are so complex – human beings are complex – and even the definition of success is complex.  There is often not a simple financial metric for success (though commercial success is also not that simple).  I will use three case studies to illustrate how we operate in an uncertain environment:

• The use of program logic and theory of change – NEWPIN program
• The use of data and proxy outcomes – On TRACC Social Impact Investment
• The use of balanced scorecards – understanding economic, wellbeing and health impacts through COVID-19.

Speaker: Ben Gales is currently Executive Director, Strategic Coordination at the Department of Premier and Cabinet (DPC). Ben has a wealth of experience in economic and social policy as well as impact investing. In his current role he is responsible for the Strategic Coordination, where his focus is on driving improved social outcomes for people in NSW. He recently moved to DPC from NSW Treasury, where he was responsible for the Economic Strategy Division, which includes the NSW Office of Social Impact Investing.

10.40-11am

Harnessing data to disrupt cycles of disadvantage

The term ‘cycles of disadvantage’ is often just a compelling idea, a narrative hook to capture the injustice with which people find themselves trapped in poverty, the odds stacked against them. But can we make cycles of disadvantage visible – empirically observable, measurable and predictable – such that they become breakable? Can we harness the potential of data to identify the strong and weak points in a cycle of disadvantage, the on-ramps through which the flows of disadvantage are increased and the off-ramps that offer people transformative opportunities to escape?

Speaker: Dr Jeni Whalan is the Chief Programs Officer at the Paul Ramsay Foundation. She has worked across sectors – government, international organisations, academia, think tanks and nonprofits – to help organisations work better for the people they serve. Prior to joining the Paul Ramsay Foundation, she led Systems Design for NSW Education, developing innovative approaches to support schools operating on the front lines of entrenched disadvantage. She has a DPhil in International Relations from the University of Oxford.

11.25-11.45am

The art and science of extracting meaningful information from electronic medical records

The widespread adoption of electronic medical records (eMR) has created unprecedented opportunities to harness information stored within eMRs to ascertain clinical outcomes and improve healthcare efficiency at large scale and low cost. However, significant challenges remain including uncertainty around data quality for secondary use of health data as well as interpreting complex data which may hold different meanings across multiple scenarios. In this presentation, we share our learnings for deriving insights from eMRs and methods for addressing uncertainty in data quality captured in eMRs to simultaneously demonstrate value to clinicians, hospital executives and ultimately patients.

Speaker: Dr Charmaine Tam is an Analytics Translator and Senior Research Fellow at Northern Clinical School and the Centre for Translational Data Science, the University of Sydney. Charmaine works at the much-needed interface between health domain experts and data engineers and scientists, ensuring that insights from the burgeoning amount of digital health data are best utilised to lead to improvements in human health within appropriate governance and privacy safeguards. On a day-to-day basis, she manages an analytics team which focuses on harnessing insights from structured and unstructured data, such as free-text and images, extracted from electronic medical record systems, working on analytics projects which align academic pursuit with health system priorities.

11.45am-12.05pm

Unprecedented times: using REMAP trials to resolve uncertainty in disease pandemics

The COVID-19 pandemic has exposed profound short-comings in the approach to generating evidence in healthcare. Despite over 4 million documented cases, and nearly 400,000 deaths, we are still little closer to identifying any management strategies that make a difference to patient outcomes. Traditional clinical trials which are designed as classical experiments are too slow, piecemeal, and ethically troublesome to be fit for the purpose of efficiently quantifying and resolving uncertainties in clinical management. A new research paradigm, the REMAP (randomised, embedded, multi-arm, pragmatic) trial, offers a new approach for achieving health systems that learn. The concept will be illustrated in relation to COVID-19.

Speaker: Professor Tom Snelling

Tom Snelling is director of the Health and Clinical Analytics team in the School of Public Health at the University of Sydney, and an infectious diseases physician in the Sydney Children’s Hospital Network. Tom is pioneering in the application of Bayesian approaches to the design, coordination, implementation and analysis of public interest studies, and is successfully leading a suite of multi-institutional collaborative learning health projects across Australia.

12.10-12.30pm

Predicting COVID-19: What did the models tell us?

At the start of the COVID-19 outbreak, a slew of models were reported nationally and internationally that covered every scenario from a best case of virtually no cases to a terrifying scenario where 100,000s of people died. The one thing that connects all of these models is simple: they were all wrong. But is that a problem? As the saying goes, “All models are wrong, some models are useful”, so while none of these models have predicted the exact future, many have been used to define our pandemic response. Ultimately, the biggest issue with pandemic modelling may not have been the models themselves, but how we as a society interpreted the uncertainty inherent in such modelling.

Speaker: Gideon Meyerowitz-Katz is an epidemiologist working in chronic disease in Sydney’s west, with a particular focus on the social determinants that control our health. He is a regular contributor with the Guardian, and writes for a range of other publications. He writes a regular health blog covering science communication, public health, and what that new study you’ve read about actually means.

12.30-12.50pm

Did the models fail?

Speaker: Professor Sally Cripps

Sally Cripps is Director of the Data Analytics for Resources and Environments (DARE) Centre. She was awarded her PhD in 2002 and her work has appeared in the world’s most prestigious statistical journals. Her research has focused on the development of new and novel methods to flexibly model and analyse complex data.

2-2.20pm

Data science with impact

In the digital era, data analytics helps industries to harness the power of data – more efficient operations, more cost savings, higher profits and happier customers. It has the capability to rapidly revolutionise traditional solutions and ways of thinking in the industry, provided that it can be followed up with real world applications and impact.

Data science discovers patterns from discriminated data, and build predictive capability from the derived patterns. The impact of data science is in utilising data to gain unique business insights, however, how to deal with and illustrate uncertainty in modelling and risks associated with probabilistic framework can be a big barrier. This talk shares insights of how to create innovative data science solutions with uncertainty measures.

Speaker: Professor Fang Chen

Fang Chen is a prominent leader in AI/data science with international reputation and industrial recognition. She is the winner the ‘Oscars’ of Australian science, 2018 Australian Museum Eureka Prize for Excellence in Data Science. In science and engineering, Professor Chen has 300+ refereed publications, including several books.

2:20-2.40pm

Bayesian optimisation for environment monitoring

In this talk Dr Marchant will explore how to make decisions under uncertainty over spatial temporal phenomena. The talk will briefly cover spatial temporal Gaussian Process regression and its use as a surrogate model in a probabilistic optimisation framework that automatically balances exploration and exploitation. The talk will also focus on theory and applications, including air pollution monitoring, autonomous safe navigation and future applications including the associated technical challenges.

Speaker: Dr Roman Marchant

Dr Marchant is a bayesian computer scientist who completed a PhD at the School of Information Technologies, University of Sydney in 2015.  His current research focuses on developing new data science methods to answer complex questions involving decision making under uncertainty.  Roman focuses on fundamental and applied areas, including social science, criminology, bushfires, air pollution and fisheries.  His area of expertise is Sequential Bayesian Optimisation (SBO), which is a novel probabilistic method for optimal sequential decision making under uncertainty that maximises long-term reward. Although SBO has been readily applied to robotics and environmental monitoring, it can be applied to any optimisation problem.

2.45-3.05pm

Business decision-making in COVID-19

Businesses are facing massive uncertainty following the lockdown and now the risk of new outbreaks with the reopening of the economy. The three biggest uncertainty factors are unemployment post stimulus, consumer behaviour and investment. Initially businesses sought priors to give some certainty in forecasts about the virus spread and economic recovery shape-V,L,U,W or J inverted but they are not helping much. Equities markets are betting on a V but most analysts think another shape more likely. Businesspeople now getting resigned to not having certainty about the recovery shape but managing whatever shape they are faced with.

Like good Bayesians with a touch of epistemic humility they are utilising an arsenal of actions consistent with continuing uncertainty-small adaptive steps linked to natural experiments, buying information about consumer and investment response from markets already opened up, lining up options for low cost equity and debt financing, taking out hedges or insurance, pursuing no regrets moves to accelerate trends such as remote working, digital transformation, rents linked to turnover and avoiding big bets for now until uncertainty lowered. These are actions to address uncertainty set out in BPS p199-201.

Speaker: Robert McLean AM is a former Senior Partner of McKinsey and Company where he led the Australian and New Zealand Practice. Rob serves as Chair of The Nature Conservancy Australia Board and as Vice Chair of the Asia Pacific Council of The Nature Conservancy, a director of the Paul Ramsay Foundation and The Centre for Independent Studies. He was the founding Chairman of Social Ventures Australia, a former President of The Benevolent Society and Dean of the Australian Graduate School of Management.

3.05-3.25pm

Data, sharing and use – what do we need to know about our data, can we actually know it, and where does it lead us?

Data sharing is often hampered by concerns related to the data itself, what it will reveal, if the insights in that “reveal” are valid, and any harms that come from the use of insights. Concerns about data and its use are often framed in terms of privacy considerations, relying on the legal frameworks of privacy legislation. The reality is however more often based on concerns about data quality, data completeness and the expertise required to fill in the gaps, or to interpret data driven insights. This presentation will look at some of the work being performed to build data sharing frameworks based on an understanding of these factors and progress towards measurement of the level of information in people centric data.

Speaker: Dr Ian Oppermann is the NSW Government’s Chief Data Scientist working within the Department of Customer Service. He is also an Industry Professor at the University of Technology Sydney.  From 2015 to 2019, Ian was also the CEO of the NSW Data Analytics Center (DAC). Ian has nearly 30 years’ experience in the ICT sector and, has led organizations with more than 300 people, delivering products and outcomes that have impacted hundreds of millions of people globally.

3.50-4.10pm

Data and decisions in natural resource stewardship

Speaker: Professor Hugh Durrant-Whyte is NSW Chief Scientist and Engineer. Hugh was Chief Scientific Advisor to the UK Ministry of Defence from 2017-18, and the founding Director of the Australian Centre for Field Robotics (ACFR) at the University of Sydney. He is a world authority on machine learning and robotics, and their application in areas including cargo handling, mining and defence.

4.10-4.30pm

Uncertainty

Much of my early career in infrastructure was focused on reducing risk and creating certainty. In recent years I have come to realise that uncertainty can be something to be valued. Some years ago there was a major step forward when infrastructure asset owners started to realise that whole of life costs need to be considered when comparing options. The trouble is, the “life” that was used in these considerations sometimes should have been a variable rather than a constant.  It may well be that I know that the “life” of a dam or a bridge can be considered to be at least a hundred years, but do I really know that I need it for a hundred years. By over-valuing large fixed assets that provide certainty we can create path dependency which locks in futures and locks out the potential for change.  Sometimes we need to accept risk for a little longer, valuing uncertainty and recognising that there is a value in keeping our options open.

Speaker: Dr Jim Bentley is Chief Executive Officer, NSW Water Sector and Deputy Secretary of WATER, NSW Department of Planning, Industry and Environment. He was formerly Managing Director of Hunter Water Corporation Newcastle NSW. Jim has extensive global experience in infrastructure services, particularly in the water industry, having spent 12 years with Thames Water, including 6 years in the international business mainly focused on Turkey and the Middle East.

Jim is Professor of Practice in Business and Law at the University of Newcastle NSW and Visiting Professor in Engineering at University College London, UK.

4.35-4.55pm

On uncertainty in mining

The mining industry generates and uses large amounts of data along the value chain, from material in ground through to any downstream processing. The information is used for a variety of purposes, including decision support and control. However, the typical representations of information in day to day operations do not explicitly maintain or quantify any uncertainty in this data. While the concept of “master data” or “one source of the truth” has been used within the industry to reduce ambiguity in information, the use of probabilistic techniques, commonly applied in other sectors such as defence and finance, are less common. However, the application of analytics and probabilistic estimation techniques to quantify and support decision making under uncertainly has potential to assist day to day operations as the industry adopts more automation technologies and data driven systems.

Speaker: Dr Eric Nettleton is an experienced innovation and technology professional who has held strategic leadership roles in multiple industries, including mining, aerospace, IT and academia. He is currently Head of Automation at Newcrest Mining, where he supports automation and technology strategy and deployments across the value chain for all Newcrest assets.
His interests in information fusion, robotics and large scale, whole-of-business technology and automation platforms aim to support significant business and productivity improvements in real world applications. Eric has a Degree in Mechatronic Engineering and a PhD in the area of information fusion and field robotics from The University of Sydney.

 4.55pm-5.15pm

Creating value from uncertainty

Speaker: Julie Batch is the Chief Strategy & Innovation Officer at Insurance Australia Group Limited (IAG), she is responsible for aligning IAG’s customer and corporate strategy and developing future growth options beyond insurance. Julie leads her division to combine functions such as customer governance, data, AI, innovation and venturing – together with strategy, building new growth pathways and developing future capabilities across Australia and New Zealand.

5.35-5.45pm

Closing remarks

Speaker: Professor Sally Cripps

Sally Cripps is Director of the Data Analytics for Resources and Environments (DARE) Centre. She was awarded her PhD in 2002 and her work has appeared in the world’s most prestigious statistical journals. Her research has focused on the development of new and novel methods to flexibly model and analyse complex data.