Dear colleagues,

This is the announcement of the CWI Scientific Meeting, with speakers Stratos Idreos and Willem Hundsdorfer, 13.00-14.00, Room Z009 (Euler).

It is a lunchtime meeting, so sandwiches will be provided before the talks.  We hope to see you there!

Best regards,
Willem Hundsdorfer and Ronald de Wolf

PROGRAM:

Date: Friday, November 30
Time: 13.00-14.00
Location: Euler Room (Z009)

Jos Baeten: Director's announcements
Lecture 1: Stratos Idreos (INS1)
     "3 Ideas for Big Data Exploration"
Lecture 2:  Willem Hundsdorfer (MAC3)
     "Numerical oscillations - and how to avoid them"

ABSTRACTS:
1. Stratos Idreos (INS1)
Today, businesses and sciences create more data than what we can store, move, let alone analyze; every two days we create as much data as much we created from the dawn of humanity up to 2003. A fundamental problem with big data is that data management systems require extensive tuning and installation steps; by the time we finish tuning, more data have already arrived. To make things worse, tuning a system requires knowledge of what to tune for, i.e., we need to know the kind of queries we will be posing. However, in several modern big data applications, we are often in need of exploring new data quickly, searching for interesting patterns without knowing a priori exactly what we are looking for. In this talk, we will discuss 3 novel ideas towards adaptive database systems which are tailored for data exploration: adaptive indexing, adaptive loading and dbTouch systems.

2. Willem Hundsdorfer (MAC3)
Numerical simulations for differential equations are often plagued by the occurrence of unwanted numerical oscillations, in particular when the exact solutions possess steep gradients or shocks. This can lead to non-physical numerical approximations, for example with negative densities or concentrations.
    The question is: what are good numerical schemes that avoid oscillations with realistic, feasible stepsizes? In this talk we will address this question for multistep methods to solve ordinary differential equations. It will be seen that the so-called implicit methods do not allow much larger stepsizes than their explicit counterparts. In practice this means that small stepsizes are unavoidable. Still, there are "good" methods and "bad" methods. The good ones allow relatively small stepsizes, the bad ones require extremely small stepsizes.