The Risk Institute Online

During the COVID-19 pandemic, we have moved our resources and events online. Follow this page to keep up to date with what's happening in the Risk Institute.

RIO TALKS - September 2022

Tuesday 06th September

Beth Montague-Hellen: LGBTQ+ Inclusion in STEM Fields

Over the last decade, STEM fields have shifted in their acceptance and inclusion of LGBTQ+ staff and students. Organisations have sprung up, awards and grants have been given out, and Universities have put on events and conferences, but is this enough? This talk will address where we are now, what the current challenges in LGBTQ+ inclusion are, and will give some thoughts on what we need to do next.

Slider 1

Dr Beth Montague-Hellen started off academic life as a Bioinformatician, but while a postdoc realised that it was much more fun supporting other people’s research. In 2016 Beth retrained as a Librarian. Since then Beth has primarily worked in Scholarly Communications and Research Support at a number of different UK Higher Education institutions and is currently a Senior Research Librarian at the University of Nottingham.

Beth started the LGBTQ+ STEM blog as a way of making LGBTQ+people working in STEM field more visible to each other and was then persuaded through the medium of twitter to organise the first LGBTQ+ STEMinar. This conference has now run for 7 years and was awarded the Royal Society Athena Prize in 2020.

RIO TALKS - July 2022

Tuesday 19th July

David Ropeik: Risk Communication is NOT (just) About the Facts

Risk is defined, generally, as the chance that something bad could happen. 'Chance' is quantifiable. 'Bad' is not. It is entirely subjective. So the challenge for risk communication is to appreciate the psychological factors that determine how a risk feels, and to demonstrate a sincere respect for those feelings in both actions and messages designed to build trust and therefore enhance the impact of the information the communicator wants to convey.

Slider 1

David P. Ropeik is a retired Harvard University Instructor, author, and international consultant and speaker on risk perception, risk communication, and risk management. He was an Instructor of risk communication at the Harvard School of Public Health, and was co-director and principal faculty member of the school’s professional education course ‘The Risk Communication Challenge’. He taught the course Critical Thinking about Environmental Issues in the Harvard School of Continuing Education from 2006-2016.
He is author of How Risky Is It, Really? Why Our Fears Don’t Always Match The Facts, (2010, McGraw Hill). He is co-author of RISK, A Practical Guide for Deciding What’s Really Safe and What’s Really Dangerous in the World Around You, (2002, Houghton Mifflin). He is author of the forthcoming Rethinking Our Fear of Cancer, How excessive worry about a dread disease does great harm all by itself.

Wednesday 13th July

Christian Schilling: A Gentle Introduction to Reachability Analysis for Dynamical Systems

In this presentation, Prof. Schilling will explain the principles of bounded-time reachability methods. These are essentially symbolic-numeric ODE solvers lifted to sets of states and can compute infinitely many trajectories at the same time. With such approaches we can exhaustively analyze systems with uncertainty in the initial condition or in the dynamics (control inputs, disturbances, etc.). An immediate application is to prove that a system cannot reach an error state under any circumstance. The presented approaches are available in the tool JuliaReach.

Slider 1

Prof. Christian Schilling received his doctoral degree in computer science from the University of Freiburg, Germany, in 2018. He was a postdoctoral research fellow at IST Austria and an interim professor for cyber-physical system at the University of Konstanz, Germany. Since 2021 he is an assistant professor in the Distributed, Embedded and Intelligent Systems group at Aalborg University in Denmark. Christian's research in the area of formal methods is focused on the analysis, verification, and synthesis of systems with dynamical or machine-learned components.

Tuesday 5th July

Silvia Tolo: Current challenges and future solutions for systems safety analysis

This talk is to be held at the Risk Institute in our Seminar room, and will be broadcast live on Zoom

This talk describes the research carried out on the Lloyd’s Register Foundation funded project NxGen (Next Generation Prediction Methodologies and Tools for System Safety Analysis) to tackle the issues of current modelling capabilities for complex system safety analysis. The research presented is investigating novel theoretical and computational strategies for overcoming the limitations of common risk assessment methodologies. An overview will be provided of the modelling framework under development, discussing its capabilities, current research challenges and future potential applications.

Slider 1

Dr Silvia Tolo gained an M.Sc. in Energy and Nuclear Engineering from the University of Bologna, and subsequently collaborated with the Institute for Risk and Uncertainty at the University of Liverpool, where she was awarded a Ph.D. She is currently undertaking research within The Resilience Engineering Research Group at the University of Nottingham on the development of theoretical and computational tools for the efficient modelling of complex systems.

RIO TALKS - June 2022

Tuesday 21st June

Dalal Alrajeh: Handling Risks to Goal Assurances in Adaptive Systems

The engineering of high-quality software requirements generally relies on assumptions about the environment in which the software-to-be has to operate. As assumptions are added or changed, so must the software requirements. Ensuring the correctness of this process is challenging and introduces risk to the safe operation of any software that implements them. In this talk, I will be presenting some of our recent work on safe-by-construction adaptations of software systems in which we try to address such problems both at design- and run-time.

Slider 1

Prof. Dalal Alrajeh is a senior lecturer (associate professor) at the Department of Computing, Imperial College London, and currently serves as the director of postgraduate research at the department. She holds a PhD in distributed software engineering from Imperial College. Her research is in the areas of formal software engineering and symbolic AI, with a focus on specification learning, synthesis and analysis for provably correct software systems. She is the recipient of several funding awards for her research, more recently for the correct design of humans in the loop, and AI-enabled systems for safety-critical applications. . She was the recipient of the Imperial College Research Fellowship award. She has served on numerous program committees of premier conferences in her field (including ICSE, EFEC/FSE, ASE, IJCAI and AAAI).

Tuesday 21st June

Dominic Balog-Way: The evolving field of risk communication: Where are we now and where are we going?

Communicating about risks to health, safety, and the environment is challenging. Although there have been notable successes, history is replete with examples of communication missteps and failures. Many well-meaning efforts have caused unintended negative effects, or, worse, boomeranged, generating the opposite effects of what was intended. Drawing on decades of interdisciplinary research, I explain why and how risk communication effectiveness would be significantly improved if practitioners adopted a more strategic and evidence-based approach. I first explain what such an approach entails, including the importance of choosing clear goals and evaluating messages throughout the process. Next, I discuss the key components of strategic risk communication that all effective practitioners must consider carefully. These are highlighted with environmental protection and public health sector examples, ranging from lead ammunition poisoning and chronic wasting disease, to smoking, underage drinking, and COVID-19. I conclude by providing concrete recommendations on how practitioners can become more strategic and resist the temptation of relying on intuition and unproven traditional practices.

Slider 1

Dr. Dominic Balog-Way is a postdoctoral research associate in the Department of Communication, Cornell University. His interdisciplinary research examines the assessment, management, and communication of risks to public health and the environment. Working internationally and across sectors ranging from environmental protection and food safety to pharmaceuticals and public health, Dominic strives to improve public policy through evidence-informed risk management and strategic benefit-risk communication. He is currently working on projects regarding the use of lead ammunition for hunting, infectious marine diseases, and deep geothermal energy. Throughout his career, he has advised, and worked closely with, state, national, and international governments, as well as businesses, advocacy groups, and academics.

RIO TALKS - May 2022

Tuesday 31st May

Enrique Miralles Dolz: Interval-Based Global Sensitivity Analysis for Epistemic Uncertainty

Epistemic uncertainties are always present at some point in the design process of complex engineering systems and evidence-based policies. Acknowledging these uncertainties and measuring their possible consequences, described with mathematical models, is of uttermost importance to increase the chances of finding a successful system or policy. Sensitivity analysis can help with this task by explaining how the input uncertainty of a mathematical model contributes to its output uncertainty.

Slider 1

Enrique Miralles Dolz joined the Institute for Risk and Uncertainty in 2018 for his doctoral studies, working on uncertainty quantification, sensitivity analysis, and optimisation under uncertainty, primarily in systems under epistemic uncertainty. Before that he studied engineering in Valencia, Spain, and physics at the York Plasma Institute.

Wednesday 25th May

Divya M. Persaud: 3D Imagery to Support Planetary Exploration

Planetary exploration is becoming more daring and pushing the boundaries of technology as we seek answers to fundamental questions of life in our solar system and the geologic history of our planetary neighbors. 3D imaging provides opportunities to mitigate risk for operating missions on other worlds, from preparing rover traverses to finding the best science targets for landers, rovers, and satellites. I will discuss some of these applications and how my work using 3D imagery of the Martian surface fits into the multi-modal approach of understanding the red planet.

Slider 1

Dr Divya M. Persaud (she/her) is a planetary geologist fascinated by the stories of rocks in the solar system. As a postdoctoral scholar at the NASA Jet Propulsion Laboratory, she is working on missions to explore Jupiter's moon, Europa. Her past work includes her doctoral research using 3D imagery to investigate the exposed rock layers at the exploration site of Curiosity rover in Gale crater, Mars, as well as working to understand the geology of the icy moons of Saturn, Mercury, and asteroids and meteorites using satellite and laboratory data. She is also a poet, composer, and speaker, and has spoken and performed her cross-disciplinary work internationally.

Tuesday 17th May

Vladimir Stepanov: Uncertainty in Agent-based Models

Researchers have found that it is nearly impossible to escape epistemic uncertainty. Thus far more models are incorporating epistemic uncertainty which many argue improve the model outputs. The KFC disruption in 2018 may have been averted if the more extreme scenarios were considered. However, agent-based models with epistemic uncertainty are underexplored. This talk will cover the current methods to deal with this uncertainty type and explore alternative methods in agent-based models.

Slider 1

Vladimir Stepanov studied BSc Computer Science at University of Edinburgh before starting on an MRes/PhD at the Risk Institute at Liverpool. In his PhD, he explores the different methods of representing epistemic uncertainty in an ABM.

Tuesday 10th May

Sara Owczarczak-Garstecka: When science has gone to the dogs: Mixed-methods approaches to studying dog welfare and human-dog interactions

There are nearly 10 million dogs in the UK and despite increase in owner's spending, at the population level, their welfare is often poor and declining. This talk will explore the main welfare issues concerning the UK dog population and discuss current and ongoing research aimed at uncovering some of these problems in depth. Particular attention will be place on showcasing how qualitative (e.g. sociological) and quantitative (e.g. statistical) analysis can be used to explore issues related to dog welfare by discussing current and ongoing studies.

Slider 1

Sara studied Anthropology at UCL and received and MSc in Clinical Animal Behaviour from the University of Lincoln before starting on an MRes/ PhD at the Risk Institute/ Institute for Infection and Global Health at Liverpool. In her PhD, she used mixed methods to explore perception and prevention of dog bites. Since completing her PhD, Sara was awarded Animal Welfare Foundation grant which enabled her to carry out post-doctoral study into access to veterinary healthcare during the COVID-19 pandemic. Sara works as a part of Dogs Trust research team.

RIO TALKS - April 2022

Tuesday 26th April

Bilal M. Ayyub: Hazard-Resilient Infrastructure Including Network Topology and Underground Spaces as Case Studies

According to the United Nations Office for Disaster Risk Reduction, the world’s inhabitants with vast majority of property and wealth are concentrated in urban centers situated in locations already prone to major disasters, such as earthquakes and severe droughts, and along the flood-prone coastlines. This concentration of the world’s inhabitants in urban centers is expected to increase from 50% in 2012 to 66% in 2025. Civil infrastructure systems traditionally have been designed, constructed, operated and maintained for appropriate probabilities of functionality, durability and safety while exposed to extremes during their full-service lives. Examining systems in the context of resilience would add proper considerations for adaptability to changing conditions including recovery. This presentation introduces a methodology for the designing hazard-resilient infrastructure, and an ASCE manual of practice that provides guidance for and contribute to the development or enhancement of standards for hazard-resilient infrastructure. The framework provided in this manual emphasizes infrastructure systems, networks and how they support community resilience. The underlying approaches are based on using probabilistic methods for risk analysis and management to address uncertainties within a planning time horizon, and are illustrated using transportation networks including metro transit and railroad systems and underground spaces.

Slider 1

Bilal M. Ayyub is an A. James Clark School of Engineering Professor and the director of the University of Maryland Center for Technology and Systems Management. He has expertise in the areas of risk, uncertainty and decision analysis and risk management including finance. He has completed research projects and studies for NSF, DOD, DOT, DHS, and several engineering companies. As a distinguished member of the ASCE, an honorary member of ASME, and a fellow of SNAME, SEI and SRA, he has served the engineering community in various leadership capacities. Among the 650 publications co-authored are more than 20 books, including several textbooks adopted for courses at universities internationally. He is the recipient of several research awards and prizes, and consulted to national and international organizations on infrastructure and defense systems on risk, microeconomic and financial modeling. He has served on the board of several research and development companies.

Tuesday 19th April

Adolphus Lye: Probabilistic Prediction of Material Properties with Artificial Intelligence (PROMAP)

In this talk, Adolphus Lye will discuss his recent work on Project PROMAP which was a feasibility study funded by the Advanced Nuclear Skills and Innovation Campus (ANSIC) as part of the Game Changers Challenge. The aforementioned study looks into addressing the issue of sparse data as well as presenting the opportunities for Artificial Intelligence (AI) within the Nuclear sector which has yet to embrace the latter.
The work presented in this session will include: 1) a brief overview of the state of AI within the Nuclear industry; 2) the method of data-enhancement on a sparse data-set; 3) the introduction of model uncertainty in Artificial Neural Networks; and 4) the amalgamation of Artificial Neural Network with Uncertainty Quantification tools. The result is a general framework to provide a robust probabilistic prediction, with the associated confidence bounds, of Nuclear material properties which is validated with experimental data.
To conclude the session, Adolphus Lye will re-iterate the benefits of the proposed methodology as well as briefly discuss the future research work to be conducted following the conclusion of this feasibility study.

Slider 1

Adolphus Lye is currently a 4th-year PhD student based at the Institute of Risk and Uncertainty within the University of Liverpool. In 2018, he graduated from the National University of Singapore (NUS) with a Bachelor of Science (with Honours) majoring in Physics and minoring in Mathematics. In that same year, he obtained the Singapore Nuclear Research and Safety Initiatives (SNRSI) scholarship which funded his PhD study. He is co-supervised by Professor Edoardo Patelli and Professor Alice Cicirello.

Through the course of his candidature, his research interest mainly revolves around the topic of Bayesian model updating and its applications in addressing Structural and Nuclear Engineering problems. This includes reviewing state-of-the-art developments in Bayesian model updating, addressing model uncertainty in inferring time-varying parameters via On-line Bayesian model updating, and more recently, merging Uncertainty Quantification tools with Artificial Intelligence algorithms to address the issue of sparse data in Nuclear engineering.

Wednesday 13th April

Nick Gray: Why Risk and Uncertainty are Key for Humane Algorithms

Algorithms have no idea about the significance of the calculations they are performing. They just mindlessly output the results of complex mathematical operations, often requiring untenable assumptions to be made, irrespective of the risk posed by even simple errors. Humane algorithms need to provide meaningful information about why a particular decision has been made and, when encountering an error, fail so that a human overseeing it can deal with the uncertainties and not increase the risk. I argue that such an algorithm needs to be able to balance the uncertainties and risks that govern even simple problems. This can be achieved by carefully considering the different types of uncertainty present and utilising the framework provided by imprecise probabilities when making calculations and analysing results.

Slider 1

Nick Gray is a PhD Student at the Institute for Risk and Uncertainty at the University of Liverpool under the supervision of Scott Ferson. His PhD studies are multidisciplinary, with research interests including machine ethics, uncertainty in machine learning, and risk communication. He recently spent six months working as a research associate at Imperial College London researching Human-in-the-loop machine learning.

Wednesday 13th April

Jeffrey Dambacher: Qualitative Mathematical Modelling

Qualitative mathematical modelling will be introduced as a means to understand and predict the dynamics of complex systems. The technique poses the question: if all we know of a system is the general nature of the relationships between species and environmental or human variables, but not the precise intensity of these interactions, then what do we know? It turns out that we know not everything, but quite a lot. Qualitative mathematical modelling describes complex systems through only the sign (0, +, -) of the effect or interaction between variables, and thus can easily include variables and processes that are important, yet difficult to measure. Calculations of system stability and predictions of perturbation response proceed through analysis of the feedback properties of a system. While model predictions are imprecise, there are nonetheless rigorously derived and readily testable. This approach leads to many interesting, practical, multidisciplinary, and surprisingly overlooked applications to complex dynamical systems. This talk will provide a brief introduction to the basic approach and include example applications and extensions to a Bayesian interpretation of results.

Slider 1

Dr Jeffrey Dambacher's background was originally in the ecology of stream fishes. A desire to understand the complex relationships of fish communities, rivers and watersheds led him to the method of qualitative mathematics as a tool to address complex systems. With CSIRO this work has been expanded to address problems of integrated monitoring and management of socio-ecological systems such as the Great Barrier Reef, marine ecosystems of Australia’s exclusive economic zone, Mediterranean fisheries and aquaculture systems in Tasmania, Chile and France.

RIO TALKS - March 2022

Wednesday 23rd March

Arianna Casanova: Information Algebras in the Theory of Imprecise Probabilities

This presentation takes up and deepens the compatibility problem; the problem of establishing if some probabilistic assessments have a common joint probabilistic model, in the framework of desirability. In particular, we prove the possibility to induce information algebras from coherent sets of gambles and coherent lower previsions, both interpreted as pieces of information about values of a group of variables. Then, we show that it is possible to obtain the same results of Miranda and Zaffalon (2020) about the compatibility problem in the unconditional case in a more simple way by using only instruments of these general algebraic structures. This allows us moreover to enforce the view of such imprecise-probability objects as algebraic and logical structures and gives tools to manipulate them as such.

Slider 1

Arianna Casanova started her Ph.D. at Dalle Molle Institute for Artificial Intelligence (IDSIA) in May 2018. Previously, she received a BSc and an MSc degree in mathematics both from the University of Milan. Her work focuses on the analysis and generalisation of the imprecise-probability formalism of desirability.

Wednesday 2nd March

Alex Winbush: Inference of diagnostic characteristics without a gold standard for disease classification

This presentation introduces a means of inferring diagnostic test statistics without a 'gold-standard' second test through possibility theory (end even without any second test at all). Precise, or imprecise, knowledge about the correlation between test results, population prevalence, and even the statistics of the comparison test can all help to improve the answer but are optional. You can even infer them as well if you have enough data!

Slider 1

Alex Wimbush is a PhD student at the Institute for Risk and Uncertainty, University of Liverpool, and is currently writing his thesis on optimising combinations of diagnostic tests under uncertainty. His work centres around confidence based methods of inference, prediction, and calculation using possibility theory and c-boxes.

RIO TALKS - February 2022

Thursday 24th February

Dr Jazmin Scarlett: The keys to the past: a mixed-methods approach to reconstructing the 1812 eruption of La Soufrière St. Vincent

The usage of a wide range of qualitative and quantitative data and integrating them in creative and compelling ways with a pragmatic underpinning, can provide deep investigations into the impacts of volcanic eruptions on society. The 1812 eruption of La Soufrière volcano on the Eastern Caribbean Island of St. Vincent was reconstructed with a mixed-methods approach in mind. This eruption occurred during the slavery era where there was a reliance on enslaved labour to cultivate the island’s sugar monoculture. Findings have found that the eruption produced ash fall, pyroclastic density currents (PDCs), volcanic earthquakes and lahars that impacted 129 plantation estates, leading to 43 documented deaths. This eruption also forced the emigration of the indigenous Kalinago and the purchasing of land for displaced enslaved Africans. Lastly, a key aspect of this eruption was that due to estate owners receiving loans, the focus was to “return back to normal” and not to adapt, essentially meaning that the dimensions of vulnerability and risk did not change as a result.

Slider 1

Dr Jazmin P. Scarlett is a Historical and Social Volcanologist, researching how people lived in the past and the present with active volcanoes. Previous research has researched the people and volcanoes of St. Vincent and the Grenadines, Montserrat, Germany and Italy. She also has interests in heritage studies, hazard analysis, disaster studies concepts and theories and science communication and outreach pedagogy. Dr Scarlett did her undergraduate in BSc (Hons) Geography and Natural Hazards at Coventry University, MSc Volcanology and Geological Hazards at Lancaster University and PhD in Earth Science at the University of Hull. She has been a visiting researcher at Aarhus University in Denmark, previously a Lecturer in Physical Geography at Newcastle University.

Tuesday 22nd February

Marco Zaffalon: Casual Inference is Imprecise Probability in Action

Causal inference is a beautiful recent theory with a wealth of important applications waiting to be pursued. It is also a native theory of imprecise probability; there is so much to be gained in creating bridges across the two fields. In this talk, I will introduce causal inference, give examples of methods and algorithms that can help you cross the bridges, and thus start right away to apply causal inference in your domain.

Slider 1

Marco Zaffalon is a University Professor and Scientific Director at IDSIA. He has 160 refereed publications on artificial intelligence and machine learning, and his research has been supported by 15 million francs in competitive grants. Marco is a Senior Area Editor of the International Journal of Approximate Reasoning and has been a past President of the Society for Imprecise Probability. At IDSIA he has founded and leads the group on probabilistic machine learning that is made of 40 researchers. In his applied research he has worked with UBS, Novartis, Mastercard, and several other companies. In 2020 he has co-founded Artificialy, a company for products and services in AI, where he works as the company's Chief Scientist.

Tuesday 15th February

Dominik Hose: Possibilistic Inference - From Imprecise Probabilities to Inferential Models

Possibility theory provides us with a useful language for describing various kinds of uncertainty. It serves not only as a mathematical description of imprecise probabilities, but it also constitutes a very natural approach to statistical inference in the framework of inferential models proposed by Martin and Liu. This talk argues in favor of these claims, it highlights the connections between the two proposed applications of possibility theory, and it demonstrates how possibilistic inferential models essentially provide the tools for efficiently computing with nested confidence intervals.

Slider 1

Dominik Hose received his B.Sc. and M.Sc. in Simulation Technology from the University of Stuttgart (Germany) in 2015 and 2017, respectively. For the last five years, he has been a PhD student working on possibilistic uncertainty quantification with imprecise probabilities under the supervision of Michael Hanss. In January 2022, he submitted his dissertation entitled "Possibilistic Reasoning with Imprecise Probabilities: Statistical Inference and Dynamic Filtering".

RIO TALKS - January 2022

Tuesday 25th January

Ekaterina Auer: Towards Assessing The Likelihood of Mutations in BRCA1/2 Genes with Interval and Dempster-Shafer Theory Based Methods

Germline mutations in BRCA1/2 genes are considered to lead to an increased risk of hereditary breast and ovarian cancer syndrome (HBOC). Modern genetic tests reliably identify BRCA1/2 mutations but are not necessarily helpful for everyone. Therefore, a preliminary step for arriving at specific suggestions concerning individual HBOC prevention and risk mitigation is the use of risk assessment tools that compute the likelihood of a mutation, for example, the Penn II risk model (https://pennmodel2.pmacs.upenn.edu/penn2/) and many others, including easy-to-use, questionnaire-type approaches. Because there are no true standards for data acquisition on the basis of which the mentioned tools decide about the mutation risk, the modeled mutation likelihoods might vary considerably so that decision uncertainty appears. The data might be incomplete wrt. patient’s origin, age, type of cancer, family history, etc. In this contribution, we take a first step towards data fusion/cleanup and propose two models to combine data on mutation probabilities for better correlation between personal and family cancer history or between different risk factors.

Slider 1

Prof. Ekaterina Auer is a professor of mathematics at the Department of Electrical Engineering of the University of Applied Sciences in Wismar, Germany, since 2015. She received her diplomas in mathematics and computer science from Ulyanovsk State University, Russia, in 2001 and from the University of Duisburg-Essen, Germany, in 2002. She worked at the Chair of Computer Graphics and Scientific Computing at the University of Duisburg-Essen as a research assistant, receiving her Ph.D. in 2007 and her postdoctoral qualification (habilitation) in 2014. Her main areas of interest are algorithms with result verification and their application to engineering problems (e.g., in biomechanics or in energy systems' simulation); uncertainty quantification and propagation using verified, stochastic, or mixed approaches; verification and validation frameworks including uncertainty visualisation; automated comparison and recommendation of verified software; and application of modern parallelization strategies (e.g., using the GPU) in the mentioned contexts.

Tuesday 18th January

Peg Coleman: Incorporating Food and Gut Microbiota into 21st Century Risk Analysis

The advances of the microbiome revolution of the past decade have deeply challenged our prior understanding of microbes in human systems biology in health and disease. There is zero uncertainty that microbes in the 21st century are now understood as symbionts ‘completing’ the human ‘superorganism‘ or ‘holobiont’ (Homo sapiens plus microbial partners in health) rather than germs that will kill us. Yet, our current frameworks for Risk Analysis exclude our microbial partners in health! This lecture will address microbes in health and disease, focusing on the gut, the gut-lung axis, and the respiratory system, as well as strategies for managing our microbes for health and protection from disease. Recently published case studies are introduced that provide evidence maps on benefit-risk analysis for mammalian milks, both fresh unprocessed (raw) and pasteurized breastmilk and cow milk. Dialogue about potential partners in the work of incorporating food and gut microbiota into 21st century Risk Analysis will close the lecture.

Slider 1

Peg Coleman, MS2, is a medical microbiologist, a microbial risk assessor, a Fellow of the Society for Risk Analysis (SRA). Ms. Coleman has served in many leadership roles in SRA since 1995 and was recently elected to serve on the SRA Council. She presented SRA webinars in 2017 and 2021 on topics related to the ‘microbiome revolution’. Ms. Coleman has served as a Mentor for graduate students and new members in the SRA mentoring project. Her long career as a microbial risk assessor began with the US federal government (USDA/FSIS) and continues as a consultant. Her primary interests are benefit-risk analysis and resilience of human superorganisms.

Wednesday 05th January

Georg Schollmeyer: Nonlinear Dynamics for Ecosystem Forecasting and Management

In many applied situations of regression analysis the variables one is actually interested in can not be directly observed or can not be observed in the resolution that is actually needed. This can for instance be due to a censoring or a coarsening of the data, or it can be due to measurement error, etc. In such situations, without further assumptions about the censoring- or coarsening process, or without additional knowledge about the measurement error, the obtained statistical model is usually only partially identified, which means that the underlying true regression parameters can not be estimated consistently. Therefore, within the methodology of partial identification, one does not try to estimate the true parameter, but instead one estimates so-called identification regions, which are subsets of the parameter space that contain all parameters that cannot be excluded with an infinite amount of observed data and the imposed model assumptions. In this talk, I would like to present certain identification regions in the context of (multiple) linear regression for the case where the outcome variable, the covariate(s), or both can only be observed in intervals. After discussing the case of interval-valued outcomes where different identification regions arise due to different imposed model assumptions, I would like to speak about the more difficult case where the covariates (and possibly also the outcomes) are interval-valued.

Slider 1

Georg Schollmeyer is a postdoctoral researcher in the working group 'Foundations of Statistics and Their Applications' headed by Thomas Augustin at the department of statistics at Ludwig-Maximilians-Universität (LMU) in munich. After studying mathematics at Technische Universität Dresden he went to munich to make a Phd in the field of imprecise probabilities. Under the supervision of Thomas Augustin he worked mostly on imprecise probabilities and partial identification in statistcs. Now, he is doing his habilitation at LMU, again under the supervision of Thomas Augustin.