Lunchtime Seminar
Archive Summer Semester 2017
Modeling Small-Scale Surface Deformations as Static Waves
Lecturer:
Evgeny Zuenko
Research assistant at IGS group, University of Innsbruck
Date: Thursday, 29th of June 2017, 12:00 – 1:00
Venue: SR 1, ICT Building, Technikerstraße 21a, 6020 Innsbruck
Abstract:
Modeling of deformation-induced surface details (wrinkles, folds, buckles, etc.) serves two main purposes – firstly, to enhance the visual realism of simulated objects; secondly, to have a measure of mechanical stresses applied to the object. This talk will introduce an approach to represent such details as a static wave over the surface of a simulated object. A short overview of related work will be given, as well as a description of the underlying physical model. The newly developed method will be introduced, followed by demonstrations of up-to-date results, and discussion about further steps.
Decoy Password Vaults: At Least as Hard as Steganography?
Lecturer:
Pascal Schöttle
Postdoctoral Researcher at SEC group,University of Innsbruck
Date: Thursday, 22th of June 2017, 12:00 – 1:00
Venue: SR 2, ICT Building, Technikerstraße 21a, 6020 Innsbruck
Abstract:
Cracking-resistant password vaults have been recently proposed with the goal of thwarting offline attacks. This requires the generation of synthetic password vaults that are statistically indistinguishable from real ones. In this work, we establish a conceptual link between this problem and steganography, where the stego objects must be undetectable among cover objects. We compare the two frameworks and highlight parallels and differences. Moreover, we transfer results obtained in the steganography literature into the context of decoy generation. Our results include the infeasibility of perfectly secure decoy vaults and the conjecture that secure decoy vaults are at least as hard to construct as secure steganography.
Heterogeneous Computing on Embedded SoCs: past and future trends
Lecturer:
Ivan Grasso
Research Assistant at DPS group, University of Innsbruck
Date: Thursday, 8th of June 2017, 12:00 – 1:00
Venue: SR 2, ICT Building, Technikerstraße 21a, 6020 Innsbruck
Abstract:
Recently, embedded computing systems have drawn a lot of attention from industry and academia due to their performance and energy advantages. Embedded CPUs and GPUs, driven by the massive growth of the phone and tablet markets are continuously evolving from a hardware and software perspective, finding new ground in several emerging application areas. In this talk, we will discuss the recent advancements in embedded systems with a particular focus on their computational capabilities. We will present the results of our research that compare embedded CPUs and GPUs, depicting advantages and drawbacks of each solution. Finally, we will conclude with future trends in the heterogeneous embedded computing field.
Watching Remote Management Protocols for Security: The Case of TR-069
Lecturer:
Maximilian Hils
Research assistant at SEC group, University of Innsbruck
Date: Thursday, 1st of June 2017, 12:00 – 1:00
Venue: SR 2, ICT Building, Technikerstraße 21a, 6020 Innsbruck
Abstract:
TR-069 is a standard for remote management of end-user devices by service providers. Despite being implemented in more than 350 million devices, almost no research has been published on the security and privacy aspects of TR-069. In this talk, we will show methods and tools to inspect TR-069 communication, and present initial findings on the security of the TR-069 ecosystem.
Context-aware music recommendations
Lecturer:
Martin Pichl
Research Assistant at DBIS group, University of Innsbruck
Date: Thursday, 18th of May 2017, 12:00 – 1:00
Venue: SR 2, ICT Building, Technikerstraße 21a, 6020 Innsbruck
Abstract:
Over the last years, music consumption has changed fundamentally: people switch from private, mostly limited music collections to huge public music collections provided by music streaming platforms. Thus, the amount of available music has increased dramatically and music streaming platforms heavily rely on recommender systems to assist users in discovering music they like. Incorporating the context of users has been shown to improve the quality of recommendations. We present a context-aware track recommender system that exploits information about the current situation and musical preferences of the user. The presented approach allows to successfully leveraging interaction effects between listening histories, situational and track content information, substantially outperforming a set of baseline methods.
On the Security and Scalability of Proof of Work Blockchains
Lecturer:
Arthur Gervais
Postdoctoral Researcher at Department of Computer Science, System Security Group, ETH Zürich
Date: Thursday, 11th of May 2017, 12:00 – 1:00
Venue: SR 2, ICT Building, Technikerstraße 21a, 6020 Innsbruck
Abstract:
The security properties of blockchain technology allow for the shifting of trust assumptions, e.g., to remove trusted third parties; they, however, create new challenges for security and scalability, which have not yet been fully understood and that we investigate in this talk. The blockchain’s security, for example, affects the ability of participants to exchange monetary value or to participate in the network communication and the consensus process. Our first contribution provides a quantitative framework to objectively compare the security and performance characteristics of Proof of Work-based blockchains under adversaries with optimal strategies. Our work allows us to increase Bitcoin’s transaction throughput by a factor of ten, given only one parameter change and without deteriorating the security of the underlying blockchain. In our second contribution, we highlight previously unconsidered impacts of the PoW blockchain’s scalability on its security and propose design modifications that are now implemented in the primary Bitcoin client. Because blockchain technology is still in its infancy, we conclude the talk with an outline of future work towards an open, scalable, privacy-preserving and decentralized blockchain.
Automatic optimization of OpenCL programs
Lecturer:
Klaus Kofler
Research Assistant at DPS group, University of Innsbruck
Date: Thursday, 4th of May 2017, 12:00 – 1:00
Venue: SR 2, ICT Building, Technikerstraße 21a, 6020 Innsbruck
Abstract:
Unleashing the full potential of GPUs is a challenging task as it requires effective parallelization strategies. The problem becomes even more challenging, when heterogeneous systems, consisting of multi-core CPUs and GPUs, are involved. This talk presents two methods, to automatically optimize the performance of OpenCL programs for GPUs and heterogeneous systems. The first method improves the performance of OpenCL kernels executed on a GPU by transforming the data layout of the program, as it is very hard for GPU programmers to identify a well-suited data layout and it has a big impact on the performance and. The second method is designed to accelerate a given OpenCL program by distributing the computational load over all OpenCL devices available. As the distribution of the workload is crucial to gain high performance, this method relies on a machine learning generated model to predict a near-optimal workload distribution a-priori.
Robot Vision for Action and Interaction
Lecturer:
Özgür Erkent
Postdoctoral Researcher at IIS group, University of Innsbruck
Date: Thursday, 27th of April 2017, 12:00 – 1:00
Venue: SR 2, ICT Building, Technikerstraße 21a, 6020 Innsbruck
Abstract:
Robots are increasing their effects across various domains of manufacturing, services, medical, space and defense. In this talk, I will discuss robot vision which is one of the key parts of controlling a robot. Robot vision must be able to adapt to different environments and its output must be robust to dynamic changes. I will present work related to pose estimation of objects for manipulation tasks, fusion of sensor information from multiple sensors, appearance-based gesture recognition for basic robot interaction and the applications of these methods on real robots.
Discriminate real and artificial human faces in digital media
Lecturer:
Giulia Boato
Aggregate Professor at Department of Information Engineering and Computer Science, University of Trento
Date: Thursday, 6th of April 2017, 12:00 – 1:00
Venue: SR 2, ICT Building, Technikerstraße 21a, 6020 Innsbruck
Abstract:
Modern computer graphics technologies brought realism in computer-generated characters, making them achieve truly natural appearance. Besides traditional virtual reality applications such as avatars, games, or cinema, these synthetic characters may be used to generate realistic fakes, which may lead to improper use of technology. This fact raises the demand for advanced tools able to discriminate real and artificial human faces in digital media. In this seminar, the state of the art on discrimination between computer generated and natural faces will be revised, discussing also recent approaches which propose solutions based on advanced computer vision technologies allowing the extraction of physiological measurements from video sequences. The analysis of such physiological signals and/or facial dynamic information can also be exploited to tackle the strictly connected problem of face anti-spoofing.
Model repositories: Requirements, state of the art & open issues
Lecturer:
Martin Häusler
Research assistant at QE group, University of Innsbruck
Date: Thursday, 30th of March 2017, 12:00 – 1:00
Venue: SR 2, ICT Building, Technikerstraße 21a, 6020 Innsbruck
Abstract:
Modeling is an important technique in a variety of fields, ranging from general software engineering to embedded systems, from the automobile industry to avionics. In some cases, models are even used dynamically at runtime. When entire teams work on large modeling projects, sharing those models and efficiently collaborating on them without proper tool support is difficult and error-prone. Model repositories aid this task by providing a platform that can manage, persist and query model data. In addition, model repositories often provide version control mechanisms and branching capabilities. In this talk, we will discuss the state of the art of model repositories, compare different solutions to each other and discuss their individual advantages and drawbacks. Afterwards, we briefly outline how we intend to improve upon this state of the art in our own research.
Smart data for behavioural change: Towards energy efficient buildings
Lecturer:
Anna Fensel
Ass.-Professor at STI, University of Innsbruck
Date: Thursday, 23rd of March 2017, 12:00 – 1:00
Venue: SR 2, ICT Building, Technikerstraße 21a, 6020 Innsbruck
Abstract:
The Semantic Web is not a separate Web but an extension of the current one, in which information is given well-defined meaning, better enabling computers and people to work in cooperation.” – this statement of Tim Berners-Lee has gained even more relevance since the start of this century. The humanity is rapidly developing and persistently experiencing local and global challenges, such as global warming/climate change, dis-balances in demand and supply, among many others. Mastering most (if not all) of them require a behavior change. Behavioral change is difficult to achieve per se, and it is important that technology – as a major enabler – has a positive rather than a negative impact here. Further, the dramatic growth of data volumes (Big Data, Internet of Things) and the data’s increased power and impact and on the people’s daily lives are calling for new types, practices and policies of behavior with data.
These factors made the role of semantic technology even more crucial: in terms of providing a well-defined meaning, and eventually delivering Smart Data for a functional and fair data value chain. Addressing the behavioural change with Smart Data, I discuss potential ICT solutions investigating the domain of energy efficient buildings. Particularly, our completed OpenFridge experiment will be presented: design and development of the Internet of Things data system with semantic and data analytics enablers for building new services on a top of typical home appliance data — in particular, refrigerators. The system has been evaluated with real life end-user pilots.
In conclusions, I overview our related ongoing work, namely, in the areas of the impact of Big Data on society and related research roadmapping (linking to sociology), personalized energy efficiency data management services in buildings (linking to psychology), and semantic data licensing (linking to law).
Certifying Complexity and Termination Proofs for Programs
Lecturer:
Rene Thiemann
Associate Professor at CL group, University of Innsbruck
Date: Thursday, 16th of March 2017, 12:00 – 1:00
Venue: SR 2, ICT Building, Technikerstraße 21a, 6020 Innsbruck
Abstract:
In recent years both the strength and the complexity of automated program analysis tools increased. Because of the latter, it regularly happens that some analysis is faulty.
In order to still maintain reliability we developed CeTA, a highly reliable and powerful certifier for (in)validating automatically generated complexity and termination proofs.
In the seminar we first present the general architecture of CeTA and then highlight some of recent challenges we had to solve for validating sophisticated complexity and termination proofs.
Certifying Complexity and Termination Proofs for Programs
Lecturer:
Benedetta Tondi
Research associate at Department of Information Engineering and Mathematics, University of Siena
Date: Thursday, 9th of March 2017, 12:00 – 1:00
Venue: 3W03, ICT Building, 2nd floor, Technikerstraße 21a, 6020 Innsbruck
Abstract:
Every day we share our personal information with digital systems which are constantly exposed to threats. Security-oriented disciplines of signal processing have then received increasing attention in the last decades: multimedia forensics, digital watermarking, biometrics, network intrusion detection, steganography and steganalysis are just a few examples.
Even though each of these fields has its own peculiarities, they all have to deal with a common problem: the presence of adversaries aiming at making the system fail. It is the purpose of Adversarial Signal Processing to lay the basis of a general theory that takes into account the impact of an adversary on the design of effective signal processing tools. By focusing on the most prominent problem of Adversarial Signal Processing, namely binary detection or Hypothesis Testing, we contribute to the above mission with a general theoretical framework for the binary detection problem in the presence of an adversary. We resort to Game Theory and Information Theory concepts to model and study the interplay between the decision function designer, a.k.a. Defender, and the adversary, a.k.a. Attacker.
We analyze different scenarios depending on the adversary’s behavior, the decision setup and the players’ knowledge about the statistical characterization of the system. Then, we apply some of the theoretical findings to specific problems in multimedia forensics: the detection of contrast enhancement and multiple JPEG compression.