Research portraits of the Anita Borg Memorial scholarship 2010 winners

Christina Pöpper

Christina's research is motivated by her abstract kind of thinking, which she has actually fostered with riddles since she was a kid, until her teacher presented her with a book on cryptography. It is a similar spirit that appeals to her in the specific field of security. She won this award with her fascination for paradoxically hidden abstract constructions and by mechanisms that are hardly graspable even as abstractions. With some of her future projects she has a “disappearing data vision” - to strengthen the privacy concerns of users by guaranteeing that expired data cannot be re-revealed. Even the comment on her web side reads like surreal lyrics: "My interests cover broad security topics from A (like Alice & Bob) to Z (like Zero-knowledge proofs)".

It is with her typical open and integrative attitude that Christina goes for research. In the group of Prof. Srdjan Capkun her current PhD work on wireless network security spans several realms, reaching from extreme abstraction, through physical experiments related to wireless communication on the hardware level, to the relevant problems of real-world applications, regarding today's wide-spread communications technologies (such as WiFi, ZigBee, mobile phone communication, GPS) as well as future visions. Before starting her PhD she worked two years in the strategic planning group for ESA, lead by her fascination for astrophysics. She gained experience in professional project management while she was developing strategic management tools and she had the chance to communicate to outstanding people such as astronauts. All her activities are energized by her personality, namely by her sound personal belief, her passion for her field, and by a dynamic lightness in the way she is able to offer and to harvest her energies.

A kind of lightness gleams even through her research interests where she physically deals with the traffic of light and energy. Her aim is to improve the availability of wireless radio transmissions by advancing anti-jamming techniques, namely the spread-spectrum techniques. These techniques make the wireless signal more robust and available for the intended receivers even under attacks and noise. The novel concept is that there is no need of pre-shared keys between the sender and the receiver before the communication takes place. Instead, the developed technique would spread the wireless signal traffic across the fences of several spectral crash barriers and hide the signal by using random distribution patterns within a broadened range of frequencies, thus making the statistical reconstruction of the signal harder and also providing resistance against interferences in wireless communication, such as noisy channels.

For a more allegoric illustration, imagine dewdrops on a spider web where each drop prism sparkles in a spectrum of colors. Now if one could use a magic watering can to spread not only water but some kind of secret tune for example of the wind, hidden in the way these colors sparkle...so they would form an orchestra of colors, where from this orchestra of sparkling blues, reds and greens and yellows one could again hear the words and meanings sent out with the breath of wind...

Andrea Francke

What Andrea finds most appealing in her scientific research, is the beautiful and challenging discovery of highly abstract thoughts and models. She loves the esthetics and elegance of abstract solutions and with this attitude she has to master rather ambitious and mathematically difficult problems. This was a strong motivation for her master thesis work in Theoretical Computer Science, done in the area of Combinatorics and Linear Programming. In the latter field, a problem is described in terms of a linear function that is to be optimized while the solution has to satisfy finitely many linear inequalities. The solution space can be represented geometrically as a space that is bounded by hyperplanes defined by the inequalities. In a way, Andrea was working with the "crystal", which remains upon rejecting whatever is beyond these constraining limits. Beyond plenty of real world applications, results from Linear Programming can also be used as building blocks in combinatorial proofs. In order to facilitate this, the elements provided for this aim need be most compact, most elegant and most general.

While trying to achieve this, there were moments when Andrea needed perseveration. Similarly as during her semester thesis, which was also in the field of Theoretical Computer Science, more precisely in Computational Geometry: She actually had gone through various proof ideas which all seemed to finish in a dead end, not leading to the desired solution. Finally, when writing up the negative results and thinking everything over, as the time allocated for her semester thesis was almost over, she suddenly discovered a gap in one of the apparent dead ends. With this, she found a way to prove the desired result and to thereby answer a question that was beforehand open for 25 years; She even succeeded to publish these results in a conference.

Andrea is currently starting her PhD in Computational Geometry and Combinatorial Optimization in the group of Prof. Emo Welzl. A recurring topic which can be used to highlight the broad field of applications these two areas have are collisions: When combining different types of data in geographic information systems in order to gain new information, one needs to detect all “collisions” or intersections between, e.g.,  rivers and roads when one wants to find all bridges in an area. Similarly, collisions are relevant in 3-D modelling of molecular shapes where one wishes to determine the areas of contact and interaction between molecule surfaces. In computer graphics, objects are modeled as simplified geometric objects like balls to allow for efficient approximate collision detection. The idea of collisions is also most relevant in models of logistic processes, where usually a large number of conditions and constraints have to be optimally met. Andrea might follow one intriguing question: How do you model optimization problems, if a certain number of constraints can be neglected up to a certain degree, i.e. where a certain number of “small” collisions are allowed? And how much more complex does this possibility make things?

Besides these well known applications of previous results from Computational Geometry, with her theoretical work she is also confronted with the situation of not knowing the context where new techniques and results might once be used - through unexpected synergies of different fields and sometimes in the far future. To have her contributions serve as useful elements in the long run, she needs to find exactly those beautifully condensed solutions that are part of her deep motivation.