Abstract: The no-arbitrage hypothesis in finance states that it is not possible to extract a risk-less gain from trading in a financial market. It's analogous to the second law of thermodynamics in physics, which states that no work can be extracted from a cyclic thermodynamics transformation. Trader can use side information to extract a gain from trading. Similarly, the Maxwell's demon construction shows that a protocol that exploits information about the microscopic state of a system can extract work from a thermodynamic cycle. Information thermodynamics has established precise bounds on the maximal work that can be produced. I shall show that, within a prototype model of a market, the same bounds apply to finance.

Reference

Information thermodynamics of financial markets: the Glosten-Milgrom model
Léo Touzo, Matteo Marsili, Don Zagier arXiv:2010.01905 (2020)

Quantum Machine Learning

29 January, 2021

Speaker: Maria Schuld, Xanadu Quantum Technologies Inc., Canada & University of KwaZulu-Natal, South Africa
Time: 14:15 CET Place: webinar at Zoom meeting

Abstract: Algorithms that run on quantum computers - so-called quantum circuits - underlie different laws of information processing than conventional computations. By optimising the physical parameters of quantum circuits, we can use them for machine learning just like neural networks. This talk highlights different aspects of quantum machine learning for near-term quantum computing, including strategies of training the quantum models with data, integration into deep-learning pipelines and ways to think about the power of quantum circuits as learners.

Low Rattling: a principle for understanding driven many-body self-organization

5 February, 2021

Speaker: Jeremy England, MIT, USA
Time: 14:15 CET Place: webinar at Zoom meeting

Abstract: Self-organization is frequently observed in active collectives, from ant rafts to molecular motor assemblies. General principles describing self-organization away from equilibrium have been challenging to identify. We offer a unifying framework that models the behavior of complex systems as largely random, while capturing their driven response properties. Such a "low-rattling principle" enables prediction and control of fine-tuned emergent properties in disordered mechanical networks, random spin glasses, and robot swarms.

Which Way Beyond the Standard Model?

12 February, 2021

Speaker: John Ellis, CERN, Switzerland & King's College London, UK
Time: 14:15 CET Place: webinar at Zoom meeting

Abstract: Elementary particles and their interactions are described very well by the Standard Model, but this leaves unresolved many open questions. Many proposed scenarios for new physics beyond the Standard Model postulate new particles and/or interactions with masses beyond the reach of current laboratory experiments such as those at CERN’s LHC. A systematic way to probe indirectly the possible existence of such particles is provided by the Standard Model Effective Field Theory (SMEFT), in which the Standard Model is supplemented by higher-dimensional effective interactions scaled by inverse powers of heavy mass scales. I will present some results from a recent search for new physics beyond the Standard Model analyzing LHC Higgs and other data within the SMEFT framework.

Coincident Event Detection in the Electron Microscope

5 March, 2021

Speaker: Jo Verbeeck, University of Antwerp, Belgium
Time: 14:15 CET Place: webinar at Zoom meeting, Meeting ID: 948 6652 0271 Passcode: Colloquia

Abstract: In this talk I will give an overview of recent developments exploiting new possibilities in single event detection in the transmission electron microscope. New hybrid pixel direct electron detectors are revolutionizing the way we detect electrons in the microscope, providing the individual localization of electrons when they hit a camera surface as well as the precise time they do so. Here we report on placing such a camera in an electron energy loss spectrometer and combining it with a digital pulse processor coupled to a 4-quadrant energy dispersive X-ray detector. This prototype setup now offers to detect all EELS and EDX events as a series of single events with time stamping. This rich dataset contains all the conventional information obtained in a more traditional EELS or EDX experiment, but now augmented with timing information that allows to correlate individual events in time. As EELS and EDX spectra are fundamentally linked to the excitation (EELS) and de-excitation (EDX) of a given atom in the sample, these events are intrinsically correlated, and this indeed shows in the experimental data. We demonstrate that this augmented dataset provides unique opportunities to remove the background in EELS spectra without any assumptions on its shape or to distinguish overlapping EDX peaks.

We believe this experiment holds great promise for applications where trace elements need to be detected, an area where traditionally EELS has had difficulties, but the event-based detection has many other benefits and can be applied to other signals in electron microscopy as well.

History of Science. Why Does It Matter?

12 March, 2021

Speaker: Annette Lykknes, NTNU, Norway
Time: 14:15 CET Place: webinar at Zoom meeting

Abstract: 2019 was declared as The International Year of the Periodic Table (IYPT) by the UN and UNESCO, as it marked the 150th anniversary of Dmitri Mendeleev’s presentation of his periodic table to the scientific community. In popular accounts, Mendeleev is often celebrated as the sole discoverer of the periodic system – even though historians have recognized several independent discoverers. Moreover, such simplified accounts ignore the complex and multifaced history of the periodic system which indeed goes beyond finding periodic patterns for organizing the elements.
As a historian of chemistry and chemistry educator, I felt compelled to take the opportunity of the IYPT to research new aspects of the history of the periodic system and to disseminate its history to the public. During the international year, I participated in a range of activities initiated by a group of educators at NTNU. I also co-authored a popular scientific book on the periodic table (with Unni Eikeseth) and co-edited two special issues of scientific journals and an anthology aimed at an international audience. Through this work, I took part in the rewriting of history, and in the dissemination of new and forgotten stories which challenged the popular heroic narratives about Mendeleev.

In this talk, I will show examples of under-communicated aspects of the history of the periodic system and discuss why history of science matters to science.

Motivation for Physics

26 March, 2021

Speaker: Maria Vetleseter Bøe, University of Oslo, Norway
Time: 14:15 CET Place: webinar at Zoom meeting

Abstract: Why do our students choose physics? How does their motivation to learn physics interact with our teaching? The talk discusses these questions in light of international and national research, and presents preliminary results from the IMPEL project, which studies motivation among physics students at five Norwegian universities, including NTNU.

Recent progress in solving some basic problems of quantum scattering in two dimensions

9 April, 2021

Speaker: Ali Mostafazadeh, Koç university, Turkey
Time: 14:15 CET Place: webinar at Zoom meeting

Abstract: Recently we have realized that the scattering of waves by an interaction potential may be identified with the dynamics of an effective nonunitary quantum system. This observation not only provides a concrete evidence for the usefulness of non-Hermitian Hamiltonian operators, but more importantly leads to a novel formulation of potential scattering with important applications in addressing some outstanding problems of scattering theory.

We present a general introduction to this formulation and discuss some of its most striking applications. In particular, we provide a characterization of potentials with identical scattering properties below a cutoff frequency, discuss the construction of potentials for which the first Born approximation is exact (a problem that was open since the publication of Born’s 1926 paper on quantum scattering), outline a method for the realization of broadband omnidirectional and unidirectional invisibility in two dimensions, and finally offer a treatment of point scatters in two dimensions that avoids the singularities of the standard methods and, hence, does not require any renormalization scheme to remove these singularities.

References

F. Loran and A. Mostafazadeh, “Exact Solution of the Two-Dimensional Scattering Problem for a Class of δ-Function Potentials Supported on Subsets of a Line,” J. Phys. A 51, 335302 (2018); arXiv: 1708.06003.
F. Loran and A. Mostafazadeh, “Potentials with identical scarreting properties below a critical energy,” J. Math. Phys. 60, 012102 (2019); arXiv: 1902.04297.
F. Loran and A. Mostafazadeh, “Exactness of the Born Approximation and Broadband Unidirectional Invisibility in Two Dimensions,” Phys. Rev. A 100, 053846 (2019); arXiv: 1904.07737.

Strong Coupling and Negative Thermophoresis

16 April, 2021

Speaker: Rodrigo de Miguel, NTNU, Norway
Time: 14:15 CET Place: webinar at Zoom meeting

Abstract: Thermal gradients may induce mass migration, even in the absence of concentration gradients. Usually, the force induced by the thermal gradient drags the particles in the direction of the heat flow, i.e. from hot to cold. This is known as positive thermophoresis, and it is commonly understood as the result of more momentum transfer from solvent particles on the hot side than on the cold side. However, various particles such as colloids, polymers, charged nanoparticles, magnetic particles, fullerenes, proteins and vesicles have been observed to migrate from cold to hot. These observations suggest that there is more to thermophoresis than plain momentum transfer resulting from collisions between hard particles.

Indeed, negative thermophoresis (drift from cold to hot) is a somewhat counterintuitive phenomenon which has thus far eluded a simple thermostatistical description. A better understanding of the mechanism behind negative thermophoresis can be important for the design and operation of advanced nanosystem properties. For example, in applications such as drug delivery, where local heating is easier than local cooling, it is desirable to guide self-propelled particles towards locally heated targets. The thermophilic motion of nanoparticles may also be used to detect and capture DNA and other biological indicators in serum, allowing for the detection of quantities otherwise unachievable in a thermally homogeneous sample. Thermophilic drift is also believed to be a key mechanism in the self-assembly of the nucleic acids and protocells which lead to the origin of life.

In this talk I present a thermodynamic framework based on the formulation of a Hamiltonian of mean force which has the descriptive ability to capture the interesting and elusive phenomenon of negative thermophoresis in an unusually elegant and straightforward fashion. When a system is strongly coupled to the heat bath, the system’s eigenenergies become effectively temperature dependent. This adjustment of the energy levels allows the system to take heat from the environment and return it as work. This effect can make the temperature dependence of the effective energy profile nonmonotonic. As a result, particles may experience a force in either direction depending on the temperature.

Quantitative Digital Microscopy with Deep Learning

23 April, 2021

Speaker: Giovanni Volpe, University of Gothenburg, Sweden
Time: 14:15 CET Place: webinar at Zoom meeting

Abstract: Video microscopy has a long history of providing insights and breakthroughs for a broad range of disciplines, from physics to biology. Image analysis to extract quantitative information from video microscopy data has traditionally relied on algorithmic approaches, which are often difficult to implement, time consuming, and computationally expensive. Recently, alternative data-driven approaches using deep learning have greatly improved quantitative digital microscopy, potentially offering automatized, accurate, and fast image analysis.

However, the combination of deep learning and video microscopy remains underutilized primarily due to the steep learning curve involved in developing custom deep-learning solutions. To overcome this issue, we introduce a software, DeepTrack 2.0, to design, train and validate deep-learning solutions for digital microscopy. We use it to exemplify how deep learning can be employed for a broad range of applications, from particle localization, tracking and characterization to cell counting and classification. Thanks to its user-friendly graphical interface, DeepTrack 2.0 can be easily customized for user-specific applications, and, thanks to its open-source object-oriented programming, it can be easily expanded to add features and functionalities, potentially introducing deep-learning-enhanced video microscopy to a far wider audience.

"Pulling rank": assessing the importance of constituents in complex organizations

7 May, 2021

Speaker: Pierpaolo Vivo, King's College London, UK
Time: 14:15 CET Place: webinar at Zoom meeting

Abstract: Assessing the relative importance of constituents in complex organizations is of primary importance in several settings. The input-output balance equation is a widely used tool to rank constituents in the most diverse scenarios: the very same tool that helps classify how species of an ecosystems or sectors of an economy interact with each other is useful to determine what sites of the World Wide Web -- or which nodes in a social network -- are the most influential. The basic principle is that constituents of a complex organization can produce outputs whose "volume" should precisely match the sum of external demand plus inputs absorbed by all other constituents to function.

The solution typically requires a case-by-case inversion of (possibly large) matrices of the 'resolvent' or 'Leontief' form -- (I - A)^{-1} -- which provides little to no insight on the structural features responsible for the hierarchical organization of resources. In this talk, I show that -- under very general conditions -- the solution of the "ranking problem" for open systems can be described by a universal master curve, which can be characterized analytically. The result follows from a stochastic formulation of the interaction matrix 'A' between constituents: using the replica method from the physics of disordered systems, the average (or typical) value of the rankings of a generic hierarchy can be computed, whose leading order is shown to be largely independent of the precise details of the system under scrutiny. We test our predictions on systems as diverse as the WWW PageRank, trophic levels of generative models of ecosystems, input-output tables of large economies, and centrality measures of Facebook pages.

Resistance Switching Mechanisms for Memory and Neuromorphic Devices

21 May, 2021

Speaker: David Gao, Nanolayers Research Computing Ltd., UK & NTNU, Norway
Time: 14:15 CET Place: webinar at Zoom meeting

Abstract: Modern computing has benefited from rapid exponential developments in both hardware speed and the availability of data but developments towards big data and machine learning techniques has fundamentally changed hardware requirements.
Nanoscale devices capable of resistance switching can be used as both memory and processing units which solves the massive bottleneck of shuffling data between storage systems and processors. Materials that switch between a single high and low resistance state can be used as extremely efficient resistance switching memory (ReRAM) devices while those that are capable of switching between multiple states can be used as neuromorphic devices (neuristors).

Despite extensive experimental and theoretical studies, the physics behind resistance switching in most materials are still poorly understood. Traditionally, qualitative physical models and mathematical formulations are employed to explain experimentally observable statistical trends but do not provide clear insight into the physical origins of these process. In order to develop and improve novel devices, we focus on the atomic scale mechanisms responsible for resistance switching across several types of materials.