Optimization is a very common process is several disciplines.
The goal of this Serendipity Session is to foster the dialogue between researchers in different fields who use variational principles, in order to foster interdisciplinary collaboration.
When: Tuesday 24 March
Where: De Salon in Huize Heyendael
Schedule:
10:00 - 10:10 Welcoming and Intro
10:10 - 10:30 First talk
10:30 - 10:40 Discussion
11:40 - 11:00 Second talk
11:00 - 11:10 Discussion
11:10 - 11:30 Coffee break
11:30 - 11:50 Third talk
11:50 - 12:00 Discussion
12:00 - 13:00 Lunch
Registration: Registration is required, please register yourself at latest 17 March via the form below. Please register yourself soon, we have only limited seats available.
Program: (session will be in English)
Speaker: Marc Slors (Professor Philosophy of Mind and Language)
Title: Stylistic cultural conventions: a synthetic philosophical approach
Abstract: For a long time critical analysis of concepts was supposed to be the core business of philosophy. Philosophers such as Eric Schliesser and Catharina Dutilh-Novaes have recently pleaded for another, complementary function of philosophy: integrating knowledge from various sciences to create a comprehensive, unified understanding of complex systems. They call this 'synthetic philosophy.' In this presentation I will present a project I have been working on for the past 6 years as an example of this approach. I will argue that sociological insights into what Bourdieu called our 'habitus', anthropological insights into culture shock, and psychological insights into 'behaviour settings' can be combined with analytical philosophy of conventions and with theorizing about cultural evolution to explain our thorough dependence on what I will call 'stylistic cultural conventions': etiquette, dress-codes, behavioural codes, architecture and the shape and design of public space. This explanation, I will argue, sheds new light on cultural frictions, for example in multicultural societies. If correct, it shows that our current way of dealing with such frictions is seriously flawed.
Speaker: Alan Sanfey (Professor Decision Neuroscience Laboratory)
Title: A decision neuroscience approach to studying human choice
Abstract: Our lives consist of a constant stream of decisions, from the mundane to the highly consequential. These decisions range from those with clearly defined probabilities and outcomes, such as how much risk to take with an investment, to those where one's outcomes depend on the choices of others, for example when playing a competitive game with an opponent. Here, I will outline our group's approach in exploring these questions, where we utilize a novel methodology to study both individual and interactive decision-making by combining the methods of behavioural experiments, functional brain imaging, and formal economic models. Examining sophisticated high-level decision behaviour at neural and computational levels can provide important clues as to the fundamental mechanisms by which decision-making operates. For example, we demonstrate that this approach reveals key motivations, other than simple economic gain, that guide our decisions in a systematic fashion. A further goal of this research program is to use the knowledge gleaned from these studies to inform public policy debates, for example in understanding how expectations play a role in financial, environmental, and health-care decisions.
Speaker: Martijn Huijnen (Professor Bioinformatics, Radboudumc)
Title: Optimization to predict disease mutations, analyze cell deformability and cluster proteins
Abstract: I will discuss three projects involving optimization approaches. The first concerns predicting whether mutations cause disease based on their location in the 3D structure in a protein. Here we used a 3D convolutional neural network to predict disease mutations and analyzed which features, besides sequence conservation, contributed to the performance of the model. I will also highlight how predictions are impacted by the underlying disease mechanisms of missense mutations. In the 2nd project we used deep learning methods to quantify the deformability of red blood cells that were infected with malaria parasites based on image data. Deep learning requires a significant amount of labeled data to train. By creating images of cells and mimicking noise and plasticity in those images, we generate synthetic data to train a network to detect and segment red blood cells from video-recordings, skipping the need for manual annotation. We uncover significant differences between the deformability of RBCs infected with different strains of Plasmodium falciparum, providing clues to the variation in virulence of these strains. In a third project we analyze migration profiles of proteins on non-denaturing gels. Such migration profiles in principle allow the detection of protein complexes, as proteins from the same complex have similar migration profiles. Nevertheless, variation between gels has to be accounted for to allow comparison of multiple experiments. We introduce using multi dimensional dynamic programming to align the gels and facilitate large scale comparisons.
