From the Delphic Oracle to weathermen and demand forecasting, mankind has used predictions to make choices. While the accuracy of prophecies was questionable, decisions in today’s fast-paced world lose their foundations before they’re even made. Yet we stand on the brink of a revolution. With AI on the rise, CEOs, politicians, doctors and even hesitant grooms are bound to reach more informed conclusions. Soon enough, faster and better prediction tools will change entire industries. And they will create new opportunities.
Building on the success of their past conferences on Blockchain and Digital Assistants, IBM Research, Swiss Re Institute, and the Gottlieb Duttweiler Institute will jointly analyze yet another fundamental evolution at “The Power of Predictions” conference:
- Technology and innovation: IBM Research will reveal the latest findings from the forefront of science. It will describe the prospects and boundaries of the new prediction technologies.
- Business: Swiss Re Institute will depict the consequences of ever better predictions for the financial services industry.
- Society: GDI will address implications for the consumer. Who will be responsible for our decisions? What happens when technology knows me better than I know myself? And finally, how will our behaviour change once we “know” the future?
Join us on 4 June 2019 at the Gottlieb Duttweiler Institute in Rüschlikon near Zurich, and meet thought leaders and decision makers from business, society, and government.
David Bosshart, CEO, Gottlieb Duttweiler Institute
Rainer Baumann, Chief Information and Digital Officer Swiss Re Management
Dr. Alessandro Curioni, IBM Fellow, VP Europe and Director IBM Research Zurich
Ajay Agrawal, Professor, University of Toronto and Founder, Creative Destruction Lab
The Power of Predictions: How Intelligent Machines Will Impact Decisions
Ran Balicer, Director, Clalit Research Institute
Predictive Care: The Doctor Will See Your Future Now
Martin Weymann, Head Sustainability, Emerging & Political Risk Management, Swiss Re Management
The End of Uncertainty: No Risk, No Insurance?
Michelle Ufford, Engineering Manager, Big Data Tools, Netflix
Gone in 60 Seconds: The Art & Science of Data-Driven Decision Making
James Bridle, artist, technologist and writer
Out of Control: When Machines Make All Decisions
GDI Gottlieb Duttweiler Institute
Predictions in Society
(This workshop session will take place in German and English.)
Health, politics, war – there is hardly an aspect of our world that will remain untouched by algorithmic predictions. These predictions can have a huge impact for individuals, society and business. Therefore, it is imperative to discuss both opportunities and risks with stakeholders before implementing predictive algorithms. Together with "Stiftung Risikodialog", the GDI created a card game to enable a discussion about opportunities and risks of predictive algorithms. Join us in playing the game and exploring possibilities to communicate and discuss the implementation of future technologies.
Jakub Samochowiec, Senior Researcher, GDI Gottlieb Duttweiler Institute
Introduction and Game Facilitator
Karin Frick, Head Think Tank, Member of the Executive Board, GDI Gottlieb Duttweiler Institute
Matthias Holenstein, CEO, Stiftung Risiko-Dialog St. Gallen
Anna-Lena Köng, project manager, Stiftung, Risiko-Dialog St.Gallen
Swiss Re Institute
Predictions in Financial Services and Beyond
(This session will take place in English.)
Evangelos Avramakis, Head Digital Ecosystems R&D, Swiss Re Institute
Digital ecosystems and predictions: The good, the bad, and the ugly
We will examine various examples in healthcare, mobility, and society and look at how these sectors take advantage of prediction techniques. We will then discuss the potential implications and limitations for the insurance industry.
Sandra Andraszewicz, Researcher and Project Leader, Behavioral Finance Team, ETH Zurich
The limits of predictability of human financial decisions
About 300 years of research in economics, psychology and mathematics aimed at developing the most accurate models that would help describe, explain and predict human financial decisions. The age of digitalization and Big Data brought a breakthrough that allowed to collect bulk data about our open and intimate behavior. However, algorithms are often blind to human intentions and unrecorded variables, as well as to extreme events. We will discuss cases, in which Big Data fails to correctly predict human behavior related to investment decisions.
Christian Klose, Senior Analytics Professional, Swiss Re Management
Is it possible to predict the next wildfire?
Societal and economic risks of wildfires have become more concerning in recent years, especially in densely populated areas and areas with industrial operations. Wildfires remain one of the least predictable perils due to their uncertainty mainly associated with wildfire ignition triggers, and the lack of knowledge about fire fuel availability, physical setting, and weather. We will look at how Deep Learning can improve the uncertainty in forecasting wildfire occurrence, severity, and temporal and spatial patterns, and how it allows decision makers to reach more informed conclusions faster by predicting wildfire prone regions several months in advance.
Moderated by Daniel Eckhart, Advocate, Swiss Re Institute
Prediction Technologies for Enterprises
(This session will take place in English.)
Dorothea Wiesmann, Leading Technologist and Department Head, Cognitive Computing and Industry Solutions, IBM Research Zurich
Building fair AI systems for enterprises
Recently, there has been a lot of discussion on biases in machine learning (ML) models and the importance of ensuring that the training data is “representative” and “unbiased”. While that is important, such a statement alone is not very actionable for practitioners who are building such models. What does “unbiased” and “representative” mean? “Unbiased” in what “scope”? “Representative” of what “scope”? Who defines that “scope”? These are all questions that we must provide answers for practitioners to build governable AI models. In this session, I’d argue that there is no such a thing as an “unbiased” ML model. So, instead of striving for unbiased ML models, it must state its biases openly.
Anika Schumann, Manager AI for Industries and Services, IBM Research Zurich
Recommender systems are pervasive in our daily lives as they became cornerstones of most internet scale applications for shopping, travel, entertainment, etc. Yet, most of these systems have been developed as black-boxes, without explaining WHY the system suggests something to a specific user. With recent developments in regulations and privacy requirements around data and AI, users now have the Right to Know and the Right to Contest any data-based decision or recommendation that they may receive. In this session, we will take a quick overview of state-of-the-art research in AI for Explainable Recommenders and highlight some of the key challenges in building and deploying them at enterprise scale.
Abdel Labbi, Distinguished Engineer & RSM, Manager AI Systems, IBM Research Zurich
Predictions are a powerful and pervasive application of narrow AI ranging from equipment failure predictions to demand predictions for fashion to predicting the onset of epileptic seizures. While in some contexts, time – both for training and inference – is not a critical aspect, in others, e.g. for medical complications in an ICU or to prevent cascading equipment damage, real-time prediction, alerting, and action is essential. In this session, we will discuss use cases as well as algorithmic and hardware advancements for significant speed-up for increasingly complex and accurate models.
Anika Schumann is an internationally recognized expert for Artificial Intelligence Diagnosis. At IBM Research – Zurich she leads a team on Artificial Intelligence for Industries that seeks to predict abnormal behavior of physical systems ahead of time.
Moderated by Chris Sciacca, Communications Manager, Global Labs, IBM Research
Martin Wikelski, Managing Director, Max Planck Institute for Ornithology
Tracking Wildlife: When Animals Predict Disasters
Thomas Ramge, journalist and author
Data Capitalism: How Artificial Intelligence Is Changing the Economy
Norbert Bolz, Professor of Media Studies, Technical University of Berlin
Artificial Intelligence: Hunting the Black Swan