18.04.2017

Thursday

04.05.2017

Prof. Markus Bär, Physikalisch Technische Bundesanstalt Berlin

Nach dem Physikstudium in die KI-Industrie oder: Wie man die schlauen Maschinen endlich zum Arbeiten bringt

Dr. Michael Köpf, Cognotekt GmbH, Köln

Die Ergebnisse jahrzehntelanger Grundlagenforschung im Bereich der Künstlichen Intelligenz schlagen sich aktuell in immer mehr Produkten und Anwendungen nieder. Mittlerweile ist KI in den Unternehmen angekommen und verändert durch die Automatisierung einfacher geistiger Tätigkeiten unsere Art zu arbeiten fundamental. Neben großen Playern wie Google, Amazon oder Baidu tragen auch mehr und mehr kleine bis mittelständische Firmen zu diesem Wandel bei. Physikerinnen und Physiker sind aufgrund der ihnen im Studium vermittelten Fähigkeiten besonders geeignet, diesen Umbruch mitzugestalten. Um ein Gefühl für dieses sehr moderne Berufsfeld zu vermitteln, werde ich anhand konkreter Beispiele aus der Versicherungsbranche die mit der Kommerzialisierung künstlicher Intelligenz verbundenen Herausforderungen vorstellen.

Einladender: Dr. O. Kamps

Thursday

18.05.2017

Vortrag im Rahmen des allgemeinen Physikalischen Kolloquiums

Experiments with freely suspended and freely floating liquid films

Prof. Ralf Stannarius, Universität Magdeburg, Institut f. Experimentelle Physik

Nonlinear dynamics of beating cilia and flagella: Swimming, steering, and synchronization

Dr. Benjamin M. Friedrich, TU Dresden, Center for Advancing Electronics Dresden (cfaed), Biological Algorithms Group

Cilia and flagella represent a best-seller of nature: their regular bending waves propel cellular swimmers such as sperm cells and green alga in a liquid. Collections of these slender cell appendages can synchronize their beat to pump fluids inside human airways and brain ventricles effectively.
In this talk, I will address the physics of flagellar swimming and how mechanical and chemical signals control these biological oscillators.
In the first part, I will a present a theory of sperm chemotaxis, i.e. the directed navigation of flagellated sperm cells in response to signaling molecules released by the egg. We show how swimming along helical paths results in an effective navigation strategy that can cope with molecular shot noise of cellular concentration measurements. Thereby, spatial information about a concentration gradient becomes encoded in the phase of an oscillatory temporal signal perceived by the swimming cell along its helical path. This theory has recently been confirmed by experiments that track swimming sperm cells in three space dimension in artificial concentration fields of signaling molecules [1].
In the second part, I will discuss flagellar synchronization as an emergent phenomenon in collections of several flagella, which arises from mutual a hydro-mechanical coupling. We present a theory of the beating flagellum as a noisy limit-cycle oscillator, which is fully calibrated by experimental data. In particular, we show using theory and experiment how external mechanical forces change speed and shape of the flagellar beat, or even stall the beat reversibly [2]. This flagellar load-response is key prerequisite for flagellar synchronization.
We present a link between the efficiency of the flagellum to convert chemical energy into mechanical work and its ability to synchronize. Finally, we characterize the beating flagellum as a noisy oscillator, whose non-equilibrium fluctuations induce stochastic phase-slips in pairs of phase-locked flagella [3].
[1] J.F. Jikeli et al.: Sperm navigation along helical paths in 3D chemoattractant landscapes, Nature Communications 6, 2015
[2] G.S. Klindt, C. Ruloff, C. Wagner, B.M. Friedrich, Load-response of the flagellar beat, Phys. Rev. Lett. 117, 2016
[3] R. Ma, G.S. Klindt, I.-H. Riedel-Kruse, F. Jülicher, B.M. Friedrich: Active phase and amplitude fluctuations of flagellar beating, Phys. Rev. Lett. 113, 2014

Einladende: Dr. S. Gurevich

Timing stability of quantum dot based semiconductor lasers

Dr. Stefan Breuer, Technische Universität Darmstadt, Angewandte Halbleiteroptik und Photonik

Passively mode-locked semiconductor lasers based on nano-scale quantum dots offer access to sub-picosecond short optical pulses thanks to their broad spectral bandwidth and ultra-fast gain dynamics. Their small footprint, multi-Gigahertz repetition rates, direct modulation capability as well as their monolithic layout makes them promising candidates for optical clock distribution, high bit-rate optical time division multiplexing and compact microwave/millimeter-wave signal generation. Their intrinsic pulse train timing stability and fixed pulse repetition rate however can be limiting factors towards their widespread implementation into time-critical applications. Experimental concepts to improve the timing stability and to enable repetition rate agility are therefore demanded. This talk will start by reviewing timing stability fundamentals of semiconductor lasers. Then, a selection of experimental concepts to improve the timing stability will be described in detail. Next, by means of a reconfigurable laser layout, higher harmonics of the fundamental repetition rate can be generated leading to a substantially improved timing stability. Finally, experimental results will be discussed and explained in the framework of a time-domain description that is able to reproduce the timing stability improvement and repetition rate agility.

Einladende: Dr. S. Gurevich

Sondertermin

On-chip generation of complex optical quantum states and their coherent control

Dr. Michael Kues, Institut national de la recherche scientifique, Varennes (Québec), Canada
Centre Energie Matériaux Télécommunications

Entangled optical quantum states are essential towards solving questions in fundamental physics, and are at the heart of applications in quantum information science [1]. For advancing the research and development of quantum technologies, practical access to the generation and manipulation of complex photon states is required. Recently, integrated (on-chip) photonics has become a leading platform for the compact, cost-efficient, and stable generation and processing of optical quantum states [2]. However, on-chip sources are currently limited to basic two-dimensional (qubit) two-photon states.
Within this presentation, I will show that integrated frequency combs (on-chip light sources with a broad spectrum of evenly-spaced frequency modes) based on high-Q nonlinear microring resonators can provide solutions towards scalable complex quantum state sources. Particularly, by using spontaneous four-wave mixing within the microring resonators, we demonstrate the generation of bi- and multi-photon entangled qubit states over a broad frequency comb spanning the telecommunications band, and control these states coherently to perform quantum interference measurements and tomographic reconstruction of their density matrix [3-5]. Moreover, we demonstrate the on-chip generation of entangled high-dimensional (quDit) states, where the photons are created in a coherent superposition of multiple pure frequency modes. In particular, we confirm the realization of a quantum system with at least one hundred dimensions. Furthermore, using off-the-shelf telecommunications components, we introduce a platform for the coherent manipulation and control of frequency-entangled quDit states.
Our results suggest that microcavity-based entangled photon states and their coherent control using accessible telecommunications infrastructure can open up new venues for reaching the processing capabilities required for meaningful quantum information science.

References
[1] Knill, E., Laflamme, R. & Milburn, G. J. “A scheme for efficient quantum computation with linear optics,” Nature 409, 46–52 (2001).
[2] Tanzilli, S. et al. “On the genesis and evolution of integrated quantum optics,” Laser Photonics Rev. 6, 115–143 (2012).
[3] C. Reimer, M. Kues, et al., “Integrated frequency comb source of heralded single photons,” Opt. Express. 22, 1023 (2014).
[4] C. Reimer, M. Kues, et al., “Cross-polarized photon-pair generation and bi-chromatically pumped optical parametric oscillation on a chip,” Nat. Commun. 6, 8236 (2015).
[5] C. Reimer, M. Kues, et al., “Generation of multiphoton entangled quantum states by means of integrated frequency combs,” Science 351, 1176 (2016).

Zeit: Mittwoch, 31. Mai 2017, 09:30 Uhr

Ort: Seminarraum 005/006 des CeNTechs 1, Heisenbergstraße 11

Einladender: Prof. Dr. C. Fallnich

Structure formation in confinement: photonic balls and active granular rotors

Prof. Dr. Michael Engel, Friedrich-Alexander Universität Erlangen-Nürnberg, Institute for Multiscale Simulation

Natural and biological systems achieve emergent behavior with elementary building blocks. Research in my group focuses on modeling structure formation processes in soft and hard condensed matter. In this talk, I discuss two joint computational-experimental works that have in common that they concern structure formation of particles in confinement. (1) A binary mixture of 3D-printed macroscopic rotors with opposite sense of rotation is excited on an electromagnetic shaker in circular confinement. Phase-separation reminiscent of spinodal decomposition is observed. Evolution of the domain size is compared to two-dimensional Langevin dynamics simulations. (2) Droplet-based microfluidics creates homogeneous emulsion droplets as sources for defined spherical confinement. We observe a discrete series of multiply twinned colloidal clusters with icosahedral symmetry. To understand and explain the formation of the clusters, we test a geometric model and extract extremal principles.

Einladender: U. Thiele

Network robustness and the impact of transmission line failures

Prof. Dr. Dirk Witthaut, Forschungszentrum Jülich GmbH, Institut für Energie- und Klimaforschung, Systemforschung und Technologische Entwicklung (IEK-STE)

The robust operation of physical distribution and supply networks is fundamental for economy, industry, and our daily life. For instance, a reliable supply of electric power fundamentally underlies the function of most of our technical infrastructure. In periods of high loads, the breakdown of a single element of the power grid can cause a global cascade of failures implying large-scale outages with potentially catastrophic consequences.
In this talk I will review the theory of transmission line outages and the robustness of networks. Once a line in the network fails, the flow must be rerouted over alternative pathways. To assess the robustness of a network we must understand where flows are rerouted and quantify how much capacity is available for this task. I will discuss the mathematical formulation of the problem, present some recent results, and demonstrate some illuminating connections to other fields of theoretical physics and applied math.

Einladender:  Dr. O. Kamps

Spontaneous and Coherent Raman microscopy for biomedical applications

Dr. Cees Otto, University of Twente, Faculty of Science and Technology, Medical Cell BioPhysics (MCBP)

Raman microscopy is well known as a powerful method for the study of molecular changes and processes. In biomedical applications the chemical information is usually extremely complex.
The laser-based implementation of vibrational spectroscopy that underlies modern spontaneous and Coherent Raman microscopy however enables a broad range of applications also in the biomedical field. In this presentation an introduction to various forms of Raman microscopy will be presented with applications on cells and tissues.

Einladender: Prof. Dr. C. Fallnich

Theory, structure and experimental justification of the metal/electrolyte interface

Dr. Manuel Landstorfer, Weierstrass Institute for Applied Analysis and Stochastics (WIAS), Berlin

Various types of modeling approaches are used to investigate the structure of the metal/electrolyte interface and its behaviour due to adsorption, intercalation and other surface effects [1]. While atomistic resolved models address reaction mechanisms, continuum models are the foundation for a comparison to voltage- or current-controlled experiments.
In this talk I will provide insight to our model framework which explicitly accounts for solvation and adsorption effects. We will show that this model is the very basis for a qualitative and quantitative understanding of the capacitive behaviour of a variety of electrolytes. We will further show that our approach is also the very basis for a model based understanding of cyclic voltammetry and thus a key tool for analytical electrochemistry.
Most common continuum models essentially relay on simplified Poisson–Boltzmann equations to model the electrochemical double layer. However, it is long known [2] that this equation is not able to predict the capacitive behaviour of electrochemical interfaces. We showed that this is due to the negligence of the incompressibility [3] and solvation effects [4] of electrolytic solutions. Our new model essentially leads to a coupled Poisson–momentum equation system and is able to predict the non-linear behaviour of the double layer capacity over the whole potential range. Additionally, our surface mixture theory [5] accounts for adsorption on the metal surface and for partial solvation, thus leading to an overall precise model which is able to predict the unsymmetric capacity curve of strongly adsorbing ions (i.e. NaF or NaClO4[5]). Based on this, we are able to predict the thermodynamic structure of the space charge layer at the metal/electrolyte interface, which is in full agreement to experimental data.

References

[1] M. Landstorfer and T. Jacob, Chem. Soc. Rev., 2013, 42, pp 3234–3252.
[2] J. Bockris, A. Reddy and M. Gamboa-Aldeco, Modern Electrochemistry 2A: Fundamentals of Electrodics, Springer, 2001, vol. 2.
[3] W. Dreyer, C. Guhlke and R. Müller, Phys. Chem. Chem. Phys., 2013, 15 , pp 7075–7086.
[4] W. Dreyer, C. Guhlke and M. Landstorfer, Electrochemistry Communications, 2014, 43 , pp 75 – 78.
[5] W. Dreyer, C. Guhlke and M. Landstorfer, Electrochimica Acta, 2016, 201, 187 – 219.
[6] G. Valette, J. of Electroanal. Chemistry and Interfacial Electrochemistry, 1981, 122 , pp 285 – 297.
[7] G. Valette, Journal of Electroanalytical Chemistry and Interfacial Electrochemistry, 1982, 138, pp 37 – 54.

Einladender: Prof. A. Heuer

Static and Dynamic Functional Brain Connectivity at Sensor and Source level: Evidences from EEG-MEG Group Analysis

Dr. Stavros Dimitriadis, Aristotle University of Thessaloniki

The human brain can be modelled as a complex networked structure with brain regions as individual nodes and their anatomical/functional links as edges. Functional brain networks are constructed by first extracting weighted connectivity matrices, and then thresholding them to minimize the noise level. Different methods have been used to estimate the dependency values between the nodes The adaptation of both bivariate (mutual information) and multivariate (Granger causality) connectivity estimators to quantify the synchronization between multichannel recordings yields a fully connected, weighted, (a)symmetric functional connectivity graph (FCG), representing the associations among all brain areas. The aforementioned procedure leads to an extremely dense network of tens up to a few hundreds of weights. Therefore, this FCG must be filtered out so that the “true” connectivity pattern can emerge.  For that reason, statistical filtering based on surrogates analysis and also topological filtering based on the maximization of information flow in the network under the constraint of the wiring cost (Dimitriadis et al., 2017) should be adopted to get a subject and condition specific functional connectivity pattern.
The whole methodology relies on data-driven techniques without using a priori information for the subjects e.g. labels. Subject-specific approaches increase the reproducibility of the dataset avoiding optimizing it independently for each study. Complementary, one can add more subjects to the original cohort without re-optimizing the parameters, an approach that can lead to controversy findings to the original study.
I will demonstrate how different connectivity estimators in both static and dynamic functional connectivity with the incorporation of arbitrary statistical and topological filtering schemes can give contradictory results. The selection of appropriate surrogates and data-driven topological filtering scheme will give more reproducible and stable results. The whole analysis will focus on electro/magneto-encephalography (EEG/MEG) at resting-state in normal populations and in mild cognitive impairments (MCI) subjects. First evidences of similarities of connectivity patterns between sensor and source-level will be also demonstrated.

References:

[1] Dimitriadis SI et al., . Topological Filtering of Dynamic Functional Brain Networks Unfolds Informative Chronnectomics: A Novel Data-Driven Thresholding Scheme Based on Orthogonal Minimal Spanning Trees (OMSTs). Front. Neuroinform., 26 April 2017 | https://doi.org/10.3389/fninf.2017.00028

Einladender: Prof. C. Wolters