Summer Semester 2018

25.04.2018 - Ankit Chauhan

Efficiency of local search on real-world networks

Many real-world networks follow structural properties like power-law degree distribution, high clustering coefficient, small-world etc. It has been observed that the very simple local search algorithms are based on searching in the k-exchange neighborhood perform very well on real-world instances. In this work, we make an attempt to understand the nice behavior of local search algorithms on real-world networks.​

02.05.2018 - Oliver Schneider

DualPanto: a Haptic Device that Enables Blind Users to Continuously Interact with Virtual Worlds

We have been developing a new haptic device that enables blind users to interact with spatial virtual environments that contain objects moving in real-time, as is the case in sports or shooter games. Users interact with DualPanto by operating its me handle with one hand and by holding on to its it handle with the other hand; these two handles are spatially registered with respect to each other. I previously presented our user study where blind participants reported very high enjoyment when using the device to play (6.5/7). We have since developed three new prototypes, two large DualPantos for higher fidelity rendering, and one mobile form factor, and developed a new software stack to support more complex applications.

09.05.2018 - Andreas Fricke

Servicification – trend or paradigm shift in geospatial data processing?

Currently, we are witnessing profound changes in the geospatial domain. Driven by recent ICT developments, such as web services, service-oriented computing or open-source software, an explosion of geodata and geospatial applications as well as rapidly growing communities of non-specialist users, the crucial issue is the provision and integration of geospatial intelligence in these rapidly changing, heterogeneous developments.

In this talk I will introduce the concept of Servicification into geospatial data processing. Its core idea is the provision of expertise through a flexible number of web-based software service modules. Selection and linkage of these services to user profiles, application tasks, data resources, or additional software allow for the compilation of flexible, time-sensitive geospatial data handling processes. Encapsulated in a string of discrete services, the approach presented here aims to provide non-specialist users with geospatial expertise required for the effective, professional solution of a defined application problem.

Providing users with geospatial intelligence in the form of web-based, modular services, is a completely different approach to geospatial data processing. This novel concept puts geospatial intelligence, made available through services encapsulating rule bases and algorithms, in the centre and at the disposal of the users regardless of their expertise.

16.05.2018 - Lung-Pan Cheng

Human Actuation

My research is about advancing immersion. Today users see and hear virtual worlds; I want users to also feel virtual worlds. Researchers have sought to better convey not only visual and auditory but also haptic feedback to enhance immersion in virtual worlds. One approach is to use mechanical equipment to provide haptic feedback, e.g. robotic arms, exoskeletons and motion platforms.

However, the size and the weight of such mechanical equipment tends to be proportional to its target’s size and weight, i.e. providing human-scale haptic feedback requires human-scale equipment, often constraining them to arcades and lab environments.

The key idea behind my research is to bypass mechanical equipment by instead leverage human power. Humans are more generic, flexible, and versatile. I thus create software systems that orchestrate humans in doing such mechanical labor—this is what I call human actuation.

23.05.2018 - Mina Rezai

Multi-Agent Generative Adversarial DomainAdaptation for Learning Multiple Clinical Tasks​

In this work, we propose a novel adversarial net-work called Radiomic-GANs for learning a joint distribution of multi-domain clinical data. We introduce multi-agent generative adversarial networks to address multiple clinical tasks end-to-end. The Radiomic-GANs comprises four components: dual generators and a couple discriminators where we consider two adaptation policy for communicating and sharing radiomic data between agents which reduce the difference between the training and test domain distributions and thus improve generalization performance. Firstly, we fix discriminators domain adaptationand only generators communicating and shared intelligence. Secondly, the generators are fixed and only intelligence sharebetween discriminators. One generator is trained on sequential multi-modal magnetic resonance images (MRI) to learn statisticaland quantitative features that results in conversion of image intomineable data and radiomic features. Second generator combines radiomic data from first generator with other patient data todevelop models for multiple clinical routine practice. Meanwhilethe discriminators are trained to distinguish generators output, coming from the ground truth or from the generator network. Our framework is generalized in the sense that it can be used in different types of clinical tasks, such as the BraTS-2017 benchmark for multiple tasks of heterogeneous tumor segmentation and prediction of patient overall survival (OS) and ACDC-2017 for tasks of cardiac MRI semantic segmentation andc ardiac disease diagnosis.


30.05.2018 - Erik Scharwächter

Low redundancy estimation of correlation matrices for time series using triangular bounds

The dramatic increase in the availability of large collectionsvof time series requires new approaches for scalable time series analysis. Correlation analysis for all pairs of time series is a fundamental first step of analysis of such data but is particularly hard for large collections of time series due to its quadratic complexity. State-of-the-art approaches focus on efficiently approximating correlations larger than a hard thresh-old or compressing fully computed correlation matrices in hindsight. In contrast, we aim at estimates for the full pairwise correlation structure without computing and storing all pairwise correlations. We introduce the novel problem of low redundancy estimation for correlation matrices to capture the complete correlation structure with as few parameters and correlation computations as possible. We propose a novel estimation algorithm that is very efficient and comes with formal approximation guarantees. Our algorithm avoids the computation of redundant blocks in the correlation matrix to drastically reduce time and space complexity of estimation. We perform an extensive empirical evaluation of our ap-proach and show that we obtain high-quality estimates with drastically reduced space requirements on a large variety of datasets.

06.06.2018 - Thomas Brand

Title tba.

Abstract tba.

13.06.2018 - Stefan Ramson

Title tba.

Abstract tba.

20.06.2018 - Julian Risch

Title tba.

Abstract tba.

27.06.2018 - Ralf Rothenberger

Title tba.

Abstract tba.

04.07.2018 - Arvind Kumar Shekar (external)

Multivariate Correlation Analysis for Supervised Feature Selection in High-Dimensional Data

In today’s scenario, several application domains involve collection of a large number of process variables also known as features. The high-dimensional feature space is commonly used for performing analytical tasks such as regression and classification. However, from the high-dimensional feature space, it is necessary to select or extract information that are relevant for a defined analytical task. The topic of multivariate correlation analysis is of paramount importance for both feature selection and extraction tasks.

The main theme of this dissertation focuses on multivariate correlation analysis on different data types. In this thesis we identify, analyze and fulfill the research gaps in the topic of multivariate correlation analysis. For this, we developed multiple novel techniques to address the correlation of the features to a target, i.e., relevance, and the correlation between the features, i.e., redundancy, on multiple data types such as continuous, categorical and time series.

Multiple views of the feature space exhibit different interactions between features and the target. Harnessing these interactions for the selection of relevant subsets may enrich the prediction model with novel information. Nevertheless, several existing feature selection algorithms focus on obtaining a single projection of the features and are not able to exploit the multiple local interactions from different feature subsets. In such datasets, few features by itself can have a small correlation with the target, but by combining these features with some other features, they can be strongly correlated with the target. Hence, it is necessary to evaluate the relevance of a feature based on its higher-order interactions in the dataset. By computing pairwise correlations, several existing works fail to address higher-order interactions between more than two features. In addition to dimensionality, the time series dataset 1 demand extraction and evaluation of a high number of subsequences for feature extraction. This requires an efficient framework to simultaneously extract relevant and novel multivariate subsequences and transform them into features. However, traditional feature transformation approaches are often unsupervised or require additional post-processing techniques.

Addressing all aforementioned problems require novel frameworks which performs large number of statistical computations. This hinders the user understanding of complex multivariate correlations. Consequently, the final problem we intend to address is enhancing the transparency of multivariate correlation analysis. Hence, in addition to the algorithmic contributions, we aim to enhance the user’s understandability of multivariate correlations in a dataset by presenting a novel software framework.

First, we present our algorithm called diverse subset selection strategy (DS3) that identifies diverse and complementary views of the dataset. We extend the concept of multiple views to our relevance and redundancy (RaR) ranking framework for mixed datasets which exhibit higher-order interactions. By evaluating the co-occurrence of patterns in multiple dimensions, our ordinal feature extraction (ordex) algorithm evaluates higher-order interactions in time series applications. Finally, we provide a software framework for exploring and understanding multivariate correlations (FEXUM), to help users understand and evaluate the multivariate correlations in the data.

In addition, this dissertation includes an extensive experimental and theoretical evaluation of the quality and scalability of our approaches with respect to the existing works. Apart from theoretical time complexity analysis, our evaluation methods are two-fold, i.e., we evaluate the proposed algorithms on synthetic and real world data. Overall, our findings show that our proposed contributions enhance the prediction accuracy and efficiency in comparison to several traditional approaches.

11.07.2018 - Vladeta Stojanovic

Title tba.

Abstract tba.

18.07.2018 - Pedro Lopes

Interactive Systems based on Electrical Muscle Stimulation

How can interactive devices connect with users in the most immediate and intimate way? This question has driven interactive computing for decades. In recent years, wearables brought computing into constant physical contact with the user’s skin. By moving closer to users devices started to perceive more of the user. allowing devices to act more personal. The main question that drives my research is: what is the next logical step? How can computing devices become even more personal?

My approach is to create devices that intentionally borrow parts of the user’s body for input and output, rather than adding more technology to the body. I call this concept “devices that overlap with the user’s body”. I’ll demonstrate my work in which I explored one specific flavor of such devices, i.e., devices that borrow the user’s muscles.

in this line of work, I created computing devices that interact with the user by reading and controlling muscle activity. My devices are based on medical-grade signal generators and electrodes attached to the user’s skin that send electrical impulses to the user’s muscles; these impulses then cause the user’s muscles to contract. While electrical muscle stimulation (EMS) devices have been used to regenerate lost motor functions in rehabilitation medicine since the ’60s, during my PhD I explored EMS as a means for creating interactive systems. My devices form two main categories: (1) Devices that allow users eyes-free access to information by means of their proprioceptive sense, such as a variable, a tool, or a plot. (2) Devices that increase immersion in virtual reality by simulating large forces, such as wind, physical impact, or walls and heavy objects.