In the scientific breakout sessions we looked towards the future: what do our scientists envisage for us in areas like artificial intelligence or quantum computing? How do we deal with data in the future? And to what level of security can you make computers safe?
14.10 BREAKOUT SESSIONS ROUND 1
* 1A: Daniel Dadush - Integer and Linear Programming Beyond the Worst-Case
I am a tenured researcher at CWI in the Networks & Optimization group. Previously, I was a Simons Postdoctoral Fellow for 2 years at the Courant Institute of Mathematical Sciences at New York University. I completed my PhD in 2012 within the ACO (Algorithms, Combinatorics, and Optimization) program at Georgia Tech under Santosh Vempala (Professor of Computer Science).
In this talk, I will explain the difficulties in obtaining rigorous explanations for the excellent practical performance of fundamental algorithms. I will then introduce the area of beyond worst-case analysis which has been developed to tackle these challenges. As a case study, I will highlight recent work in this area on the simplex method for linear programming and branch and bound for integer programming, two fundamental techniques in optimization.
presentation slides* 1B: Marten van Dijk - Secure Smart Low Lands
I am the group leader of the Computer Security group at CWI. I am a computer security researcher who investigates and develops new techniques targeting solutions of foundational security problems. My aim is to bring rigorous cryptographic thinking to security engineering.
The attack surface of a computing system/environment is the collection of attack points to which an adversary has access. How can we obfuscate or eliminate attack points? How can we reason about the attack surface of a large scale infrastructure such as a smart low lands?
no slides available
* 1C: Benjamin Sanderse - Structure! Predicting the future with models and data
Benjamin Sanderse works as a tenured researcher in the Scientific Computing group, focusing on numerical methods for uncertainty quantification and for solving partial differential equations occurring in fluid flow problems.
Machine learning techniques have become immensely popular for image recognition, playing games, and much more. To extend their use to physical systems, such as for climate predictions, we propose to embed "structure" in machine learning methods. By embedding physical laws into neural networks, we are able to "learn more with less": networks that generalize better, are more robust, and need less data.
presentation slides* 1D: Irene Viola - Quality of experience for immersive communication
I am a tenure track researcher in the Distributed and Interactive Systems group. My interests lay in multimedia compression, transmission, and quality evaluation. My research is currently focused on Quality of Experience (QoE) metrics and methodologies for immersive multimedia systems (i.e., volumetric video for XR applications). In Dec 2020 I was awarded a NWO Women In Science Excel (WISE) grant,
Advances in telecommunication systems and innovative network solutions indicate that transmission of immersive media contents will become more widespread in the future, as bandwidth capacity becomes progressively larger. Measuring the QoE in immersive scenarios allows us to understand which factors play a role in the human perception of multidimensional contents, building models that take into account user behaviour and navigation.
Such models will help optimize acquisition, transmission and rendering in an end-to-end system considering all its components, and fine-tuning each part in order to achieve the best possible results.
15.25 BREAKOUT SESSIONS ROUND 2
* 2A: Peter Bosman - Toward Evolutionary Intelligence
Peter Bosman is a senior researcher in the Life Sciences and Health group. Dr. Bosman's fundamental research focus is on the design and application of evolutionary algorithms (EAs) for single- and multi-objective optimization, and machine learning.
If natural evolution is capable of achieving natural intelligence, it stands to reason that artificial evolution should be capable of achieving artificial intelligence. Concordantly, in this talk, I present a general notion and long-term vision of Evolutionary Intelligence. I also present current-day state-of-the-art practice and possibilities, including an example of real-world societal impact that we have been able to make at CWI through an application in medicine that led to clinical uptake at our partner institute Amsterdam UMC: automatically designing internal radiation treatment plans (brachytherapy) for prostate cancer.
presentation slides* 2B: Stacey Jeffery - Quantum Software Copy Protection
I've been a Senior Researcher in the Algorithms and Complexity group since January 2017, where I hold an NWO WISE Fellowship, and a Veni Grant. My main areas of interest are quantum algorithms and cryptographic protocols, and models of quantum computation.
Quantum information cannot be copied, leading to intriguing possibilities in cryptography that would not be possible in a classical world. One of these is software copy protection. I will talk about some recent theoretical steps towards this long-term goal.
presentation slides* 2C: Lisa Kohl - Secure Computation with Silent Preprocessing
Since October 2020, I am a researcher in the Cryptology Group. My main field of interests are cryptography, coding theory and complexity theory. A special focus of my work lies in exploring new directions in secure computation with the goal of developing practical post-quantum secure protocols.
Secure computation allows two or more parties to jointly compute on their data without compromising individual privacy. In this talk, I give a high-level introduction to secure computation and present a road towards more practical secure computation, where the major part of the preprocessing phase can be performed silently, that is without communication between the parties involved.
presentation slides* 2D: Hannes Muehleisen - Saving the Planet one Query at a Time
I'm a senior researcher in the Database Architectures group.
Big data needs big clusters, or does it? Higher data processing efficiency allows single-node scalability of data analysis tasks, greatly reduces their environmental footprint.