Speaker Talks, Abstracts & Biographies
Food security to 2040: What’s coming down the track?
This presentation initially discusses the two-way interactions between human activities related to food and nutrition security (FNS: in the context of the ‘food system’), and environmental change (in the context of ‘planetary boundaries’). It will also introduce some of the impacts of crossing these boundaries for FNS. The presentation will then cover factors that determine consumption, and the health and environment outcomes of possible future consumption patterns. These will be discussed in relation to the multiple roles of the ‘food chain’ (alternatively referred to as the food ‘value chain’), and the quantity and quality of food produced. This will be linked to broader food system sustainability metrics related to social, environmental and business aspects, discussed in relation to their relevance to different stages of the ‘food chain’, and the varying ‘world views’ of the varied food system actors. Improving understanding food system governance arrangements among these actors will be central to increasing food system sustainability.
Bio: Trained in soil science, John Ingram gained extensive experience in the 1980s working in Africa and Asia in agriculture and forestry research projects. In 1991 John was recruited by NERC to help organise, coordinate and synthesise research on global change and agroecology, part of IGBP’s international global change research programme. In 2001 he was appointed the Executive Officer for the international research project "Global Environmental Change and Food Systems" (GECAFS), then based in NERC’s CEH Wallingford site. The GECAFS International Project Office was relocated to ECI in October 2006. On the close of GECAFS in 2011 he was appointed NERC Food Security Leader until assuming his current role at the ECI in May 2013.
From Food Poverty to Food Banks: Examples of Successful Failures!
In 2000, a study noted two food banks in Britain, but warned that we may very well find that food banks become as prevalent here as they are elsewhere. These predictions have proved accurate in the UK there are now close to 1000 food banks in every corner of the country, with more opening each week. So why have they grown in numbers? And more importantly for this paper why have they become an accepted part of the social landscape, popular among politicians, the public and certain sections of the media? To criticise food banks and what they offer is to risk public approbation.
Siebel’s theory of successful failures the concept is applied to food banks? Siebel’s belief was that in certain circumstances organisations are believed to be doing the best they could in difficult conditions and their voluntary or faith based nature made criticism socially unacceptable. There is a halo effect around such organisations doing ‘good’. He noted that organisations tasked with helping the disabled find work were widely celebrated across society, even though they found very little actual work for any disabled people. Such organisations not only survived, but thrived with resources flooding in year after year.
The fact that the public may feel reforms are unfair and contrary to their core values of social justice will create electoral risk, and will therefore be avoided by policy makers. Food banks and emergency food aid provision by civil society and faith based groups show something is being done while the broader welfare reforms often continue apace. Using evidence from Europe, the USA, Canada and Australia the case will be explored for food banks as successful failures.
Bio: Martin Caraher is professor of food and health policy at the Centre for Food Policy at City University London. He has worked on issues related to food poverty, cooking skills, local sustainable food supplies, the role of markets and co-ops in promoting health, farmers markets, food deserts & food access, retail concentration and globalisation. Current research interests include:
• The role of local and regional food and food projects in promoting health.
• The planning process nationwide in promoting food projects and promoting access to healthy foods.
• Government/industry interfaces in public health. Work includes the development of the English responsibility deal; the Australian and Irish national food plans.
• Cooking skills among young people, the changing nature of food skills and the culinary transition.
• The role of cooking projects and celebrity chefs in helping change dietary habits.
• The role of food markets in promoting health and well-being.
• The food supply chain and ethics.
Martin has contributed to books on public health and health promotion, including a chapter on international public health in the Oxford Handbook of Public Health. He recently edited a special edition of the British Food Journal on food banks (2014, Vol, 116(9)).
He was a founder member of the London Food Board which advises the Mayor of London on food in London and was the public health representative on the London 2012 Olympic Food Advisory Board. Consultancies include work for the UK Dept of Health, the World Bank and the World Health Organisation. He sat on the National Institute for Health and Clinical Excellence (NICE) advisory board on preventing CVD and is a member of two scientific committees 1) the Irish Government’s - safefood and 2) the International Obesity Taskforce . Current research involves collaborations with researchers in Australia, Portugal and the US. Martin was the Australian Healthways fellow for 2008. In 2012 and 2013 he was the ‘Thinker in Residence’ at Deakin University, Melbourne.
Household food insecurity and food bank usage: what we know from Canada and the UK and key knowledge gaps
This presentation will bring together research from Canadian and UK contexts to highlight what we have learned and key knowledge gaps about household food insecurity. Firstly, I will share insights afforded by the regular measurement and monitoring of household food insecurity in Canada juxtaposed against information on food bank usage in the population. The number and characteristics of people using food banks suggest that this population is a non-representative subset of the food insecure population. I will highlight how who goes to food banks is a function of the nature of food bank operations and the severity of food insecurity. Next, I will present my research on the rise of Trussell Trust Foodbanks in the UK, which has examined where food banks have opened and are most used in relation to local area socio-demographic characteristics and recent welfare reforms. Lastly, I will highlight key aspects of the body of work from Canada relevant to the UK context, and the knowledge gaps in both countries for informing future prevention- and intervention-focused food insecurity research.
Bio: I joined the Department of Sociology at the University of Oxford as a postdoctoral researcher in February 2014. My work focuses on exploring links between social policy, nutrition and health. Specifically, I have been focusing on examining welfare reform and local authority spending cuts in relation to the rise of Trussell Trust Foodbanks and homelessness. I am also conducting cross-country analyses of food hardship in the EU.
Before moving to the UK, I completed my PhD in Nutrition from the University of Toronto. My research focused on household food insecurity and food bank use in Canada, including analyses of survey data from 500 low income families in Toronto to characterize drivers of food insecurity, co-occurring material hardships, and the effectiveness of current charitable responses to the problem. Other work in the Canadian context has included examination of provincial poverty reduction strategies in relation to food insecurity.
Moving farming and food systems towards farmer well-being, consumer food security and ecosystem sustainability – what policies and practices could make a difference glocally?
The complexity of systems of food production, distribution and consumption at multiple relevant scales has long been a daunting challenge for those located in particular organizational and geographic locations who seek to understand the systems and bring about change. Yet a plethora of thoughtful analyses, social movements, innovative practices, and courageous policies have been underway over the last two decades e.g. nutrition sensitive agriculture, food sovereignty, global governance - with important public health implications. Drawing on examples from research on shifting mixed farming-food systems in Andean South America, highland and urban East Africa, and rural-urban Ontario, this presentation will outline a set of constraints and opportunities for moving systems towards greater public and ecosystem health, from the local to the global.
Bio: Donald Cole trained as a physician and practiced primary care, public health, occupational health and environmental health in a variety of settings in Canada and lower and middle income countries. A Tri-Council Eco-Research fellowship in environmental epidemiology led to research on environmental contaminants, ecosystems, human nutrition, and health. With the International Potato Center and support from the CIDA, IDRC and the Bill and Melinda Gates Foundation, he has co-led research on agriculture, nutrition and food security in South America and East Africa. He has acted as a consultant to the public health branch of the Ontario Ministry of Health, has been Acting Medical Officer of Health in Grey-Bruce, and advises other Ontario public health units. He is a member of the research-evaluation group of the Canadian Community of Practice in Ecosystem Health. As a tenured Professor at the Dalla Lana School of Public Health, University of Toronto, he teaches, mentors, and contributes research evidence to public health practice and policy globally. Within the groups identified for the workshop, his expertise would be in the food domain, with a focus on how the food system impacts on farmer, farmworker and consumer health.
Making the best of ‘variation’
Food production is dependent on living organisms and their responses to their environment. ‘Variation’ in these living organisms and their environment offers both opportunities and challenges to farmers. This presentation will consider some of the ways that farmers exploit and cope with variation and highlight opportunities that increased knowledge about natural systems, together with new technology, may offer for future food production capability.
Bio: Rosemary is a crop scientist with interdisciplinary interests in food. As an applied entomologist working closely with the fresh produce industry she has had a long-standing interest in methods of reducing the environmental impact of crop production, whilst maintaining crop yield and quality. She has achieved this particularly through the development and refinement of the components of Integrated Pest Management Strategies for several crops, including the development of weather-based pest forecasts that are used widely by the industry. Rosemary’s considerable interest in the impact of weather on the development of pest infestations has also led to collaborative studies on the impact of climate change on agricultural production systems. In recent years Rosemary has undertaken interdisciplinary research on food in collaboration with colleagues based at Warwick and elsewhere and she is one of the Academic Leads for the University of Warwick’s Global Research Priority on Food. She is a Member of the UK Advisory Committee for Releases to the Environment, The Royal Horticultural Society Science Committee, the UK Insecticide Resistance Action Group and the International Organisation for Biological Control – West Palaearctic Regional Section Council.
Taxing unhealthy food may increase social inequalities in nutrition
The aim of the present study was to explore the impact of food price policies on the expenditures and nutritional quality of the food baskets chosen by low- and medium-income households.
Experimental economics was used to examine two price manipulations: i) a fruit and vegetable price subsidy named “fruit and vegetables condition”; ii) a healthy-product subsidy coupled with an unhealthy-product tax named “nutrient profile condition”. The nutrient profiling SAIN,LIM system was used to classify each individual food according to its overall nutritional quality which then allows for a food item to be taxed or subsidized. Women from low- (n = 95) and medium-incomes (n = 33) selected a daily food basket, first, at current prices and then at manipulated prices. The extent of savings induced by subsidies and of costs generated by the tax where compared according to the income groups. Energy density (kcal/100 g), free sugars (% energy) and the mean adequacy ratio (MAR) were used as nutritional quality indicators.
Before price manipulations, low-income women selected less expensive and less healthy baskets than medium-income ones. After price manipulations the nutritional quality improved for both income group (energy density decreased, MAR increased), but savings and costs effects were less favourable for low-income women and their nutritional quality improvements were significantly lower.
Low-income women derived fewer financial and nutritional benefits from implemented food subsidies and taxes than medium-income women, suggesting that food price policies may improve diet quality while increasing socio-economic inequalities in nutrition. This study is now published: Int J Behav Nutr Phys Act 2014; 11:66.
Bio: Nicole Darmon works at the INRA within the NORT MRU in Marseille. She coordinates research aimed at assessing the relevance of existing dietary recommendations and at developing new approaches to enhance the nutritional adequacy within the concept of sustainable diets, which embraces nutrition/health, environmental, social and economic dimensions. Starting from a base of knowledge about nutritional requirements, data on consumption and food preferences of individuals, and data on product characteristics (nutrients and bioavailability of nutrients, contaminants, price, environmental impact, cultural acceptability and availability, convenience ...), the team has adopted innovative methods to model and experiment different ways of improving the nutritional quality of foods and food choices: nutrient profiling of food, diet modeling under constraints, interventions in real-life conditions.
Food scares: How can we reduce the impacts?
The increased industrialisation of food production systems has resulted in long, complex supply chains which often involve many actors across multiple countries. Food scares are now common, and cause economic losses, social disruption and wasted resources. There is a wide variety of types: for example, incidents of microbiological contamination include bovine spongiform encephalopathy (BSE), campylobacter and listeria, while physical contamination incidents include the presence of carcinogens and foreign bodies. However, the 2013 horse meat scandal was arguably the incident that has caused the greatest policy interest in food scares during recent times. This incident raised concerns about the increasing potential for food fraud and ‘food crime’, and resulted in the UK Government commissioning a review into the integrity and assurance of food supply networks.
As with any complex system, prevention of food scares is not straightforward and it is particularly challenging as there is frequently a lack of transparency along supply chains, with many actors not having full knowledge of their supply chains. This is a specific problem in the food industry as the year round demand of consumers generally does not take account of the highly seasonal nature of many ingredients, and this results in frequently changing supply chains.
In this talk I will describe work undertaken as part of the Evolution and Resilience of Industrial Ecosystems (ERIE) programme based at the University of Surrey, funded by the EPSRC as part of the ‘Complexity Science in the Real World’ initiative. This includes development of an agent based model to explore how improved information flows along food supply chains can reduce waste caused by food scares. A feature of the project is the involvement of stakeholders in collaborative model development, and the benefits of this approach will be discussed.
Bio: Angela Druckman’s research focuses on sustainable consumption and production, taking a multi-disciplinary, whole-systems thinking approach. Angela read Engineering at Cambridge University and achieved Chartered Engineer status through working in research and development in the field of electronics. She has had a wide variety of experience, including lecturing in electronics, commercial sector management, and non-governmental advocacy on climate change. She is currently Senior Lecturer in Sustainable Energy & Climate Change Mitigation at the Centre for Environmental Strategy, University of Surrey.
'Impact of food insecurity on diet'
The United States Department of Agriculture’s definition of food insecurity is: ‘limited or uncertain availability of nutritionally accepted or safe foods’. In developed countries it is not enough to have access to food but food that contributes to a balanced, healthy diet. Focusing on results from the UK and international surveys, this presentation will look at the relationship between poverty and food choices and the consequences in terms of obesity and micronutrient deficiencies. It will also consider findings from surveys that have measured food insecurity quantitatively. The data highlights that even within food insecure households themselves, some members are at greater risk of inadequate dietary intakes.
Bio: Caireen Roberts is a Senior Researcher in the Health Policy team at NatCen Social Research and a registered nutritionist. Since joining NatCen in 2006 Caireen has worked as part of the collaborative team running the National Diet and Nutrition Survey. She has extensive experience of quantitative analysis of dietary data and has contributed to many of the NDNS reports. Caireen started her research career at King’s College London and worked on the Low Income Diet and Nutrition Survey. Her main areas of interest are food poverty and diet in the elderly.
Title: Bayesian data assimilation for vector borne disease response: Theileria orientalis (Ikeda) in New Zealand cattle
In August 2012, the first case of the cattle blood parasite Theileria orientalis (Ikeda) was detected in a herd in northern New Zealand. This organism, transmitted between cattle by a tick vector, represented a new strain from which New Zealand was previously free. Experience in Japan and Australia had shown that whereas endemic strains of Theileria posed no particular threat to the cattle industry, the Ikeda strain caused significant morbidity and mortality in affected herds. By the end of 2014, there were 1200 detected infections in herds across the North Island. This talk will present work commissioned by the Ministry for Primary Industries to understand the drivers for the spread of this disease, and to predict its likely extent. The unique problem faced in this project was that although the location of cattle and the animal movement network between farms was well recorded, little was known about the national scale ecology of the tick vector. To address this, a Bayesian data assimilation approach was constructed, in which vector presence was modelled as a discrete-space latent process with a continuous-time seasonality. A joint likelihood function assimilates the epidemic data with results from a national surveillance programme designed for a different though related disease. A spatiotemporal inhomogeneous Poisson process is used to model the epidemic, with an a priori independent hierarchical binomial surveillance model. This joint model was fitted to ongoing case detection data using a non-centered trans-dimensional MCMC algorithm, integrating over the marginal posterior of the latent vector surface, censored herd infection times, and the presence of undetected infections. The algorithm was implemented using GPGPU technology to accelerate within-chain likelihood calculations to an overnight timeframe. Through the Bayes predictive distribution, real-time forecasts of disease spread were made throughout 2013 and 2014 for decision support purposes.
Bio: I work at the interface between computational statistics and epidemic management. I wrote my doctoral thesis on "Real-time Inference and Risk-Prediction for Notifiable Disease of Animals", and continued this research as a postdoctoral fellow at the University of Warwick.
My work focuses on designing fast, efficient Markov chain Monte Carlo algorithms to fit complex dynamical models to epidemic data in a Bayesian context. Accurate parameter estimation in these models is essential for accurate risk forecasting, and the quantitative measurement of statistical uncertainty inherent in the epidemic process. I use a variety of statistical techniques such as non-centred parameterisation and adaptive MCMC, and computational techniques such as distributed parallel computing and general purpose GPUs.
If that sounds complex, then another arm of my research is involved in making complex statistical techniques accessible to researchers. Perhaps more importantly, I am also interested in employing visualisation techniques to make the results of these risk forecasts accessible to policy makers for ongoing decision making during a disease outbreak
Emulator-based inference for models of large-scale infectious disease systems
Statistical inference for mechanistic models of infectious disease spread is often very computationally expensive. Such models are generally fitted in a Bayesian Markov chain Monte Carlo (MCMC) framework, which requires multiple calculation of what is often a computationally cumbersome likelihood function. This problem is especially severe when there are large numbers of latent variables to compute. Here, we propose a method of inference based on so-called emulation techniques. Once again, the method is set in a Bayesian MCMC context, but avoids calculation of the computationally expensive likelihood function by replacing it with a Gaussian process approximation of the likelihood function built from simulated data. We show that such a method can be used to infer the model parameters and underlying characteristics of spatial disease systems, and that this can be done in a computationally efficient manner.
Bio: Rob Deardon is an Associate Professor of Biostatistics with a joint position in the Faculty of Veterinary Medicine and Department of Mathematics & Statistics at the University of Calgary. Most of his recent work has been in the area of infectious disease modelling, but he is also interested in Bayesian & computational statistics, experimental design, disease surveillance methods, spatio-temporal modelling, and statistical modelling in general. He currently has a research group of over 10 graduate students. Previous to his post at Calgary he had a 8-year long faculty position at the University of Guelph and postdoctoral positions at the Universities of Cambridge and Warwick.
Looking at data and public health problems with Bayesian glasses
First, this talk states the Bayesian paradigm. Then, there will be a discussion of how the Bayesian lens is used to look at data and how to interpret and synthesis this information to understand complex models. Also, there will be a discussion of the inherent blind spots of the Bayesian paradigm and the vital role that social science theories play in understanding complex issues. This talk will close by asking some general questions which will hopefully move forward the project on evidence based decision supports for food security.
Bio: Michael Escobar is biostatistician and a professor at the Dalla Lana School of Public Health at the University of Toronto. He works on both the development of new statistical methodology and on applications in statistics as a collaborator in public health and biomedical research. His statistical methodological work has mostly been in nonparametric Bayesian methods. As a statistical collaborator, he has published papers which have looked at 11 of the 12 determinates of health.
Bayesian decision networks for risk assessment and decision support
Bayesian decision networks (BDNs) have become a state-of-the-art modelling technology to support decision-making under uncertainty. Their key features include: (1) representation of the underlying causal process, in the nodes and arcs in the graphical model, whose visualisation improves understanding and acceptance; (2) quantification of the uncertain relationships between variables in the model using conditional probability distributions; (3) the ability to capture and combine both objective information (e.g. from data or the literature) and subjective information (e.g. expert opinion); (4) powerful methods to compute new probability estimates as new evidence becomes available; (5) providing both diagnostic and predictive reasoning; and (6) the ability to combine probabilities with costs/benefits in a rigorous decision-theoretic framework to support human decision making. I will give examples of how BDNs have been used for areas such as biosecurity risk assessment and ecological modelling, which are aspects of the broader food security challenge. I will also describe how so-called Object-Oriented BDNs can be used to do more complex modelling by: combining simpler component models in a hierarchical manner; integrating BDNs with GIS systems to do explicit spatial modelling; representing change over time. Together these features will help us to build models to support decision-making about when, where and how to take actions to achieve desired outcomes or improvements in complex real-world problems such as food security.
Bio: Professor Ann E. Nicholson is the Associate Dean Education in the Faculty of Information Technology. Ann specialises in Bayesian network modelling and is an expert in dynamic Bayesian networks, planning under uncertainty, user modelling, Bayesian inference methods, and knowledge engineering with Bayesian networks. Ann is also a partner in Bayesian Intelligence which provides a wide range of consulting services in the use of Bayesian network technology.
The bycatch of Bayes nets: expanding BN inferences for improved environmental decision-making
Bayesian Networks (BNs) provide a useful statistical framework for modelling complex environmental systems. They can also provide a wealth of inferences that can facilitate improved decision-making for a wide range of users. In this presentation, I will describe three BNs that we have developed in collaboration with external partners, some of the issues involved in the development and quantification of the models, and the expanded understanding of the respective systems that the BNs have enabled. The first example is a systems approach for plant pest risk analysis in international trade, developed with funding from the WTO and in partnership with Imperial College London and five Southeast Asian countries. The second example is a BN for identifying optimal ecological windows for dredging in harbour marine environments, in collaboration with the Australian Institute for Marine Sciences, which interfaces with broader BNs developed to produce so-called scorecards for the health of the harbour.
Bio: Kerrie Mengersen is a Research Professor in Statistics at Queensland University of Technology in Australia. She is also a Deputy Director of the Australian Research Council Centre of Excellence in Mathematical and Statistical Frontiers for Big Data, Big Models and New Insights, which has three Collaborative Environments: Healthy People, Sustainable Environments and Prosperous Societies. Kerrie’s methodological interests are in Complex Systems modelling, Bayesian statistics, mixture models, hierarchical modelling and meta-analysis. Her applied interests are in biometrics, biostatistics, environmetrics and industrial statistics.
European models and software for exposure and risk-benefit assessments: food consumers, agricultural workers and residents/bystanders
I will give an overview of several projects carried out by Fera and EU partners for intake and risk assessment models related to food. The main aim has been to extend probabilistic modelling within the EU, to integrate this within established software tools, and to encourage the quantification of uncertainty as part of the risk assessment process. The work started with a focus on single pesticides, but has been extended to account for multiple pesticides, other contaminants and nutrients, and consideration of health effects (balancing risks and benefits). The main projects covered will be:
Acropolis: A web-based system and databases to predict aggregate and cumulative exposure to pesticides. Aggregate exposures take into account dietary and non-dietary sources of exposure. Cumulative exposure represents exposure to groups of pesticides within common assessment groups, with a range of RPFs (relative potency factors).
TDS-Exposure: TDS measures pooled samples of different food groups as-consumed. They also account for the impact of cooking practices and other forms of processing. TDS are efficient means of investigating total intakes of a wide range of contaminants and nutrients. The drawback is that information is lost as a result of pooling. The project will extend MCRA for TDS exposure modelling and investigate the impact of uncertainty and variability using additional data sources.
Browse: Pesticide exposures to bystanders, residents, operators and workers. Probabilistic models Outputs from Browse provide an example of non-dietary exposure to be combined with dietary exposures within Acropolis.
Qalibra: Considered the balance of risks and benefits from consumption of oily fish and other foods. A general web-based system was developed which requires input data on DALY/QALYs (disability/quality adjusted life years) and distributions of for each of the possible health outcomes, as well alternative dietary intake scenarios for the population.
I will also mention related projects and collaborations aiming to build on and integrate these components, and remaining challenges. Software will be demonstrated.
Bio: Dr Marc Kennedy: PhD from the University of Nottingham (Bayesian statistics). Research positions at Nottingham, Sheffield and the National Institute of Statistical Sciences. Since 2006 I have worked at the Food and Environment Research Agency (an executive agency of Defra), on a series of EU projects to improve the modelling of dietary and occupational exposures to pesticides for regulators. European partners have developed new web-based systems which attempt to harmonise European databases and probabilistic models.
Probability for future weather: how close are we to decision actionable expert judgement?
Whether you are interested in policy support for food security, energy planning, water management or the planning of future buildings and cities, there is a set of numbers that you would really like to have: A probability distribution for future weather, or at least part of one. It is easy to see how these numbers could fit in to any decision making framework for any of the above examples and would be incredibly useful, perhaps even necessary (e.g. for forcing models). In this talk I will explore what numbers we do have and whether or not they are worth having. We will explore different strands of data in climate change prediction and the different approaches to policy support within that field. I will close with some thoughts on how we could obtain numbers worth having, what they would look like and what this might imply for down stream decision support (e.g. in food security).
Bio: I am a Bayesian statistician with a principal interest in Bayesian modelling, uncertainty quantification and decision making for complex physical systems. In many diverse areas of science, complex physical systems are studied through the use of computer models. I am interested in how one draws inference regarding the past and future behaviour of the actual system by combining model runs, past observations of the system and expert judgement.
Quantifying the uncertainty in climate change predictions is a high-profile example of a discipline in which each of these strands of evidence (climate models, climate data and expert judgement) must be combined, and I am active in this application area. In particular, I am interested in developing methodology for building emulators (fast statistical representations of computer models that predict the computer model output at untried settings of the inputs and report our uncertainty on those predictions) and using them as part of the overarching uncertainty quantifications.
Uncertainty quantification is important, usually because decision makers require information to make decisions under uncertainty. I am interested in how one combines information from computer models, data and expert judgement in order to provide policy support for decision makers, and my PhD thesis was written on this subject.
I am also interested in what uncertainty statements mean. As a subjective Bayesian, I view uncertainty statements as judgements that must belong to someone in order to have meaning. I am therefore interested in prior elicitation, the role of the expert in statistical modelling and in internal and external analyses for Bayesian calculations.
Statistical modelling of volcanoes, with some reflections on expert judgement and uncertainty
I describe ongoing work on a first-stage statistical risk assessment for active volcanoes, in collaboration with two volcanologists. Who has learnt what during this collaboration? And how does this kind of collaboration affect the outcome of expert elicitation exercises?
Bio: My research interests are model-based inference for complex systems: combining observational data, evaluations of models, and expert judgements. How do we quantify notions of model-imperfection? How do we combine information from several different models of the same system? How do we accommodate very long model-evaluation times? My main application area is climate prediction, notably in response to transient forcing from greenhouse gas emissions. An important strand is promoting an understanding of uncertainty and probability in science and public policy.
Causality and consistency in networked decision support systems
In recent years Bayesian models increasingly need to accommodate diverse subjective judgments from a variety of different panels of experts with very different domain knowledge. These judgments about the different features of a problem then need to be drawn together into a unified coherent picture of the composite. The talk will examine how and when this might be formally possible and explore when such a composite, in a way plausible to users of the system and the outside world, can be presented. The inferential challenges of building such a system as it applied to decision support to help local and national government address food poverty will then be explained.
Bio: My interests span foundational, methodological and applied Bayesian statistics and decision theory often in high dimensions. I am currently holder of an EPSRC award to work with Liz Dowler and Rosemary Collier to investigate ways groups of experts can ensure the coherence of their judgments when managing food crises. I have spent many years studying the theory of graphical models, and Bayes Nets especially in developing dynamic and causal variants of these. Over the last few years I have also been developing a new graphical tool - more expressive than the discrete Bayes Net - called a Chain Event Graph. I am currently investigating various dynamic and decision theoretic variants of these with Ann Nicholson, Peter Thwaites and Bob Cowell. I have also begun a new collaboration with academics at Leicester University to investigate how these methods might be used in development of decision support in self-diagnosis of late onset diabetes.
Feeding the world: getting the data right for decision-making
The increasing demand for food puts pressure on social, economic, and ecological systems while simultaneously increases volatility and risk. Without change, the probable consequences are enormous, both for the environment and humans. Professor Molly Jahn will examine how to improve decision-making on the complex issue of feeding the world, looking at collaborative work aimed at improving knowledge systems that better reflect dynamics, risk, resilience and uncertainty related to food, water, energy and climate.
Bio: Professor Molly Jahn leads a global alliance of research organizations focused on building and testing modern knowledge systems for sustainability. At the University of Wisconsin, she has been named a Wisconsin Institutes of Discovery Fellow, and holds appointments in the Department of Agronomy, the Center for Sustainability and the Global Environment and the Global Health Institute. She holds adjunct appointments as senior faculty or research scientist at the US Department of Energy’s Oak Ridge National Laboratory, Columbia University’s Earth Institute and the University of Oxford’s Martin School. She has previously served as dean of the University of Wisconsin’s College of Agricultural and Life Sciences and Director of the Wisconsin Agricultural Experiment Station. In 2009-10, she was called to provide interim leadership as Deputy and Acting Under Secretary of Research, Education and Economics at the U.S. Department of Agriculture. She consults globally for business, governments, philanthropic organizations, and international multi-lateral institutions focused on agriculture, food, water and energy security, systemic risks originating in food system failures, life sciences and environment.
Andy Davis & Jonathan Horsfield
‘Feeding Britain’ – A Shock for Middle England?
Whether it be implementing central government policy or reacting to national and global events, local government is at the sharp end of service delivery to local people. At the same time, perceptions of hunger and good nutrition as an issue for local government are relatively recent, and awareness needs to continue to grow. It was from these starting points that Warwickshire County Council was asked if it wished to collaborate with Warwick University's research around developing a decision support system.
This presentation will explore the changing range, complexity, and context within which local service delivery takes place, along with key points that have emerged over the last twelve months. We will focus on the practical application of decision making at the local level using support systems. We will also highlight the complexity of decision making in smaller geographical areas between partners with differing priorities and drivers underpinning local public service delivery.
In painting the picture of typical local public services partnership arrangements that exist in local areas across the UK, the presentation will explore specific requirements that make up the evidence base for decision making where there is also a reliance on the use of support systems.
The presentation will suggest that if the support model is to be relevant and widely used, then it needs to be designed to be low cost, highly accessible to end users and to actively engage with those setting and influencing the national level decisions within which much of local delivery will need to operate. We will further highlight the opportunities for building closer relationships between the academic community and local support systems decision makers, and highlight how our experience could be replicated in other local areas across the country.
We see the clear benefits of having a better evidence base to support decisions in relation to a food crisis. We also see the potential for the model for supporting forecasting in relation to the objectives of other local services, not necessarily linked to food and hunger.
Bios: Jonathan is Renewable Energy Adviser with Warwickshire County Council. In his career he has worked in economic development, urban and rural regeneration, environmental sustainability (including climate change adaptation and sustainable construction). He currently focuses on renewable energy projects on the Council’s estate and it was this that led him to a greater understanding of the global energy picture, the potential for future scarcity and challenges for the delivery of local public services and local economies. It was through that work and questions asked about the linkages between energy and food prices – later on led to Warwickshire CC being invited to collaborate with Warwick University in its research around decision support and a future UK food crisis.
Andy is the Manager of the Warwickshire Observatory at Warwickshire County Council. The Observatory provides a research and intelligence service for the County Council and works with other local public services in the county to encourage better use of evidence to inform decision making. Andy has been with the Observatory since it was established in 2006, and prior to that led on research and intelligence for the planning, transport, development, regeneration, and economic development functions of the council. Andy was also Chair of LARIA (the Local Area Research & Intelligence Association), a professional UK wide support organisation for local researchers in public services, between 2007 and 2012. He is also currently the Treasurer of LARIA.
Barriers to Participation in Child Nutrition Programs in America
Federal child nutrition programs form the first line of defense against child food insecurity in America. Barriers to program access form a web of logistical, programmatic, socio-economic, and cultural challenges that necessitate an interdisciplinary modeling approach. Using the Summer Meals Program as a case study, we articulate and quantify the barriers to program access, with the aim of identifying systematic obstacles that prevent food-insecure children and their families from using available nutrition resources.
Bio: Rachel Wilkerson currently lives in Waco, Texas, where she works with the research team at Texas Hunger Initiative and teaches data science for Baylor University’s graduate students in Economics. A native of Lubbock and a Baylor grad herself, she also holds a Master’s in Complexity Science from the University of Warwick in England and has spent time studying and working in Budapest and Amsterdam. She aims to understand the complex systems that undergird a multifaceted problem like food insecurity by helping to bridge the gap between data-driven research and human needs on the streets.
Towards UK food security decision support
In this talk I will discuss the progress made to date in forming a decision support system for policymakers responsible for UK food security and describe the additional challenges we have found along the way.
Bio: Martine is funded by EPSRC to work on coherent inference over a network of probabilistic systems for decision support with applications to food security with Prof. Jim Q Smith. Following an undergraduate degree in mathematics, Martine joined Warwick’s Centre for Complexity Science for her MSc and PhD. Her thesis is entitled "What is the added value of using non-linear models to explore complex healthcare datasets?" and explored methods for understanding differences in patient responses to treatment in order to support medical decision-making. Martine first joined the Department of Statistics in 2012 to work with Professor Jane Hutton, Professor Jim Smith and Professor Tony Marson, exploring the role of Chain Event Graphs in modelling and managing chronic diseases, before embarking on the current project in 2013.
