How does the mind organize thoughts? The hippocampal-entorhinal complex is thought to support domain-general representation and processing of structural knowledge of arbitrary state, feature and concept spaces. In particular, it enables the formation of cognitive maps, and navigation on these maps, thereby broadly contributing to cognition. It has been proposed that the concept of multi-scale successor representations provides an explanation of the underlying computations performed by place and grid cells. Here, we present a neural network based approach to learn such representations, and its application to different scenarios: a spatial exploration task based on supervised learning, a spatial navigation task based on reinforcement learning, and a non-spatial task where linguistic constructions have to be inferred by observing sample sentences. In all scenarios, the neural network correctly learns and approximates the underlying structure by building successor representations. Furthermore, the resulting neural firing patterns are strikingly similar to experimentally observed place and grid cell firing patterns. We conclude that cognitive maps and neural network-based successor representations of structured knowledge provide a promising way to overcome some of the short comings of deep learning towards artificial general intelligence.

This data curation work introduces the first large-scale dataset of radial k-space and DICOM data for breast DCE-MRI acquired in diagnostic breast MRI exams. Our dataset includes case-level labels indicating patient age, menopause status, lesion status (negative, benign, and malignant), and lesion type for each case. The public availability of this dataset and accompanying reconstruction code will support research and development of fast and quantitative breast image reconstruction and machine learning methods.

This study investigates the internal representations of verb-particle combinations, called multi-word verbs, within transformer-based large language models (LLMs), specifically examining how these models capture lexical and syntactic properties at different neural network layers. Using the BERT architecture, we analyze the representations of its layers for two different verb-particle constructions: phrasal verbs like 'give up' and prepositional verbs like 'look at'. Our methodology includes training probing classifiers on the internal representations to classify these categories at both word and sentence levels. The results indicate that the model's middle layers achieve the highest classification accuracies. To further analyze the nature of these distinctions, we conduct a data separability test using the Generalized Discrimination Value (GDV). While GDV results show weak linear separability between the two verb types, probing classifiers still achieve high accuracy, suggesting that representations of these linguistic categories may be non-linearly separable. This aligns with previous research indicating that linguistic distinctions in neural networks are not always encoded in a linearly separable manner. These findings computationally support usage-based claims on the representation of verb-particle constructions and highlight the complex interaction between neural network architectures and linguistic structures.

Polarizers are key components in optical science and technology. Thus, understanding the action of a polarizer beyond oversimplifying approximations is crucial. In this work, we study the interaction of a polarizing interface with an obliquely incident wave experimentally. To this end, a set of Mueller matrices is acquired employing a novel procedure robust against experimental imperfections. We connect our observation to a geometric model, useful to predict the effect of polarizers on complex light fields.

Photonic integrated circuits (PICs) play a pivotal role in many applications. Particularly powerful are circuits based on meshes of reconfigurable Mach-Zehnder interferometers as they enable active processing of light. Various possibilities exist to get light into such circuits. Sampling an electromagnetic field distribution with a carefully designed free-space interface is one of them. Here, a reconfigurable PIC is used to optically sample and process free-space beams so as to implement a spatially resolving detector of amplitudes and phases. In order to perform measurements of this kind we develop and experimentally implement a versatile method for the calibration and operation of such integrated photonics based detectors. Our technique works in a wide parameter range, even when running the chip off the design wavelength. Amplitude, phase and polarization sensitive measurements are of enormous importance in modern science and technology, providing a vast range of applications for such detectors.

We present a simple method to obtain an upper bound on the achievable secret key rate in quantum key distribution (QKD) protocols that use only unidirectional classical communication during the public-discussion phase. This method is based on a necessary precondition for one-way secret key distillation; the legitimate users need to prove that there exists no quantum state having a symmetric extension that is compatible with the available measurements results. The main advantage of the obtained upper bound is that it can be formulated as a semidefinite program, which can be efficiently solved. We illustrate our results by analysing two well-known qubit-based QKD protocols: the four-state protocol and the six-state protocol. Recent results by Renner et al., Phys. Rev. A 72, 012332 (2005), also show that the given precondition is only necessary but not sufficient for unidirectional secret key distillation.

In this article we consider the numerical modeling and simulation via the phase field approach of two-phase flows of different densities and viscosities in superposed fluid and porous layers. The model consists of the Cahn-Hilliard-Navier-Stokes equations in the free flow region and the Cahn-Hilliard-Darcy equations in porous media that are coupled by seven domain interface boundary conditions. We show that the coupled model satisfies an energy law. Based on the ideas of pressure stabilization and artificial compressibility, we propose an unconditionally stable time stepping method that decouples the computation of the phase field variable, the velocity and pressure of free flow, the velocity and pressure of porous media, hence significantly reduces the computational cost. The energy stability of the scheme effected with the finite element spatial discretization is rigorously established. We verify numerically that our schemes are convergent and energy-law preserving. Ample numerical experiments are performed to illustrate the features of two-phase flows in the coupled free flow and porous media setting.

We develop a theoretical model for polarization-selective phonon pumping induced by magnon-phonon coupling in a ferromagnetic/non-magnetic acoustic bilayer structure, focusing on the effects arising from a misalignment between the magnetic and crystallographic symmetry axes. Our model considers the coupled equations of motion describing uniform magnetization dynamics (the Kittel mode) and elastic waves in both layers, incorporating phonon pumping and boundary conditions at the interface. We show that even small misalignments lift the degeneracy of transverse shear elastic modes, resulting in phononic birefringence characterized by distinct propagation velocities for linearly polarized modes. Furthermore, our analysis reveals that magnon-phonon hybridization gives magnetic-field-dependent properties to otherwise non-magnetic phonons. We show that the polarization transfer between linearly polarized phonons and the circularly polarized Kittel mode can be tuned with an external magnetic field. Our theoretical results quantitatively reproduce recent experimental findings [1].

This article presents a systematic quantitative performance analysis for large finite element computations on extreme scale computing systems. Three parallel iterative solvers for the Stokes system, discretized by low order tetrahedral elements, are compared with respect to their numerical efficiency and their scalability running on up to $786\,432$ parallel threads. A genuine multigrid method for the saddle point system using an Uzawa-type smoother provides the best overall performance with respect to memory consumption and time-to-solution. The largest system solved on a Blue Gene/Q system has more than ten trillion ($1.1 \cdot 10 ^{13}$) unknowns and requires about 13 minutes compute time. Despite the matrix free and highly optimized implementation, the memory requirement for the solution vector and the auxiliary vectors is about 200 TByte. Brandt's notion of "textbook multigrid efficiency" is employed to study the algorithmic performance of iterative solvers. A recent extension of this paradigm to "parallel textbook multigrid efficiency" makes it possible to assess also the efficiency of parallel iterative solvers for a given hardware architecture in absolute terms. The efficiency of the method is demonstrated for simulating incompressible fluid flow in a pipe filled with spherical obstacles.

We introduce a novel concept for motion robust optical 3D-sensing. The concept is based on multi-line triangulation. The aim is to evaluate a large number of projected lines (high data density) in a large measurement volume with high precision. Implementing all those three attributes at the same time allows for the "perfect" real-time 3D movie camera (our long term goal). The key problem towards this goal is ambiguous line indexing: we will demonstrate that the necessary information for unique line indexing can be acquired by two synchronized cameras and a back projection scheme. The introduced concept preserves high lateral resolution, since the lines are as narrow as the sampling theorem allows, no spatial bandwidth is consumed by encoding of the lines. In principle, the distance uncertainty is only limited by shot noise and coherent noise. The concept can be also advantageously implemented with a hand-guided sensor and real-time registration, for a complete and dense 3D-acquisition of complicated scenes.

We demonstrate that electric-dipole scatterers can mimic chiral light-matter interaction by generating far-field circular polarization upon scattering, even though the optical chirality of the incident field as well as that of the scattered light is zero. The presented effect originates from the fact that electric-dipole scatterers respond selectively only to the incident electric field, which eventually results in depolarization of the transmitted beam and in generation of far-field circular polarization. To experimentally demonstrate this effect we utilize a cylindrical vector beam with spiral polarization and a spherical gold nanoparticle positioned on the optical axis -- the axis of rotational symmetry of the system. Our experiment and a simple theoretical model address the fundamentals of duality symmetry in optics and chiral light-matter interactions, accentuating their richness and ubiquity yet in highly symmetric configurations.

Digital individual participant data (IPD) from clinical trials are increasingly distributed for potential scientific reuse. The identification of available IPD, however, requires interpretations of textual data-sharing statements (DSS) in large databases. Recent advancements in computational linguistics include pre-trained language models that promise to simplify the implementation of effective classifiers based on textual inputs. In a subset of 5,000 textual DSS from ClinicalTrials.gov, we evaluate how well classifiers based on domain-specific pre-trained language models reproduce original availability categories as well as manually annotated labels. Typical metrics indicate that classifiers that predicted manual annotations outperformed those that learned to output the original availability categories. This suggests that the textual DSS descriptions contain applicable information that the availability categories do not, and that such classifiers could thus aid the automatic identification of available IPD in large trial databases.

Representation theorems for formal systems often take the form of an inductive translation that satisfies certain invariants, which are proved inductively. Theory morphisms and logical relations are common patterns of such inductive constructions. They allow representing the translation and the proofs of the invariants as a set of translation rules, corresponding to the cases of the inductions. Importantly, establishing the invariants is reduced to checking a finite set of, typically decidable, statements. Therefore, in a framework supporting theory morphisms and logical relations, translations that fit one of these patterns become much easier to formalize and to verify. The $\lambda\Pi$-calculus modulo rewriting is a logical framework designed for representing and translating between formal systems that has previously not systematically supported such patterns. In this paper, we extend it with theory morphisms and logical relations. We apply these to define and verify invariants for a number of translations between formal systems. In doing so, we identify some best practices that enable us to obtain elegant novel formalizations of some challenging translations, in particular type erasure translations from typed to untyped languages.

Precise 3D measurements of rigid surfaces are desired in many fields of application like quality control or surgery. Often, views from all around the object have to be acquired for a full 3D description of the object surface. We present a sensor principle called "Flying Triangulation" which avoids an elaborate "stop-and-go" procedure. It combines a low-cost classical light-section sensor with an algorithmic pipeline. A hand-guided sensor captures a continuous movie of 3D views while being moved around the object. The views are automatically aligned and the acquired 3D model is displayed in real time. In contrast to most existing sensors no bandwidth is wasted for spatial or temporal encoding of the projected lines. Nor is an expensive color camera necessary for 3D acquisition. The achievable measurement uncertainty and lateral resolution of the generated 3D data is merely limited by physics. An alternating projection of vertical and horizontal lines guarantees the existence of corresponding points in successive 3D views. This enables a precise registration without surface interpolation. For registration, a variant of the iterative closest point algorithm - adapted to the specific nature of our 3D views - is introduced. Furthermore, data reduction and smoothing without losing lateral resolution as well as the acquisition and mapping of a color texture is presented. The precision and applicability of the sensor is demonstrated by simulation and measurement results.

We present the quantum programming language cQPL which is an extended version of QPL [P. Selinger, Math. Struct. in Comp. Sci. 14(4):527-586, 2004]. It is capable of quantum communication and it can be used to formulate all possible quantum algorithms. Additionally, it possesses a denotational semantics based on a partial order of superoperators and uses fixed points on a generalised Hilbert space to formalise (in addition to all standard features expected from a quantum programming language) the exchange of classical and quantum data between an arbitrary number of participants. Additionally, we present the implementation of a cQPL compiler which generates code for a quantum simulator.

Photon number resolving detectors can enhance the performance of many practical quantum cryptographic setups. In this paper, we employ a simple method to estimate the statistics provided by such a photon number resolving detector using only a threshold detector together with a variable attenuator. This idea is similar in spirit to that of the decoy state technique, and is specially suited for those scenarios where only a few parameters of the photon number statistics of the incoming signals have to be estimated. As an illustration of the potential applicability of the method in quantum communication protocols, we use it to prove security of an entanglement based quantum key distribution scheme with an untrusted source without the need of a squash model and by solely using this extra idea. In this sense, this detector decoy method can be seen as a different conceptual approach to adapt a single photon security proof to its physical, full optical implementation. We show that in this scenario the legitimate users can now even discard the double click events from the raw key data without compromising the security of the scheme, and we present simulations on the performance of the BB84 and the 6-state quantum key distribution protocols.

A Quantum Key Distribution (QKD) network is an infrastructure capable of performing long-distance and high-rate secret key agreement with information-theoretic security. In this paper we study security properties of QKD networks based on trusted repeater nodes. Such networks can already be deployed, based on current technology. We present an example of a trusted repeater QKD network, developed within the SECOQC project. The main focus is put on the study of secure key agreement over a trusted repeater QKD network, when some nodes are corrupted. We propose an original method, able to ensure the authenticity and privacy of the generated secret keys.

Stencil computations consume a major part of runtime in many scientific simulation codes. As prototypes for this class of algorithms we consider the iterative Jacobi and Gauss-Seidel smoothers and aim at highly efficient parallel implementations for cache-based multicore architectures. Temporal cache blocking is a known advanced optimization technique, which can reduce the pressure on the memory bus significantly. We apply and refine this optimization for a recently presented temporal blocking strategy designed to explicitly utilize multicore characteristics. Especially for the case of Gauss-Seidel smoothers we show that simultaneous multi-threading (SMT) can yield substantial performance improvements for our optimized algorithm.

The mass of the Galactic dark matter halo is under vivid discussion. A recent study by Xue et al. (2008, ApJ, 684, 1143) revised the Galactic halo mass downward by a factor of ~2 relative to previous work, based on the line-of-sight velocity distribution of ~2400 blue horizontal-branch (BHB) halo stars. The observations were interpreted in a statistical approach using cosmological galaxy formation simulations, as only four of the 6D phase-space coordinates were determined. Here we concentrate on a close investigation of the stars with highest negative radial velocity from that sample. For one star, SDSSJ153935.67+023909.8 (J1539+0239 for short), we succeed in measuring a significant proper motion, i.e. full phase-space information is obtained. We confirm the star to be a Population II BHB star from an independent quantitative analysis of the SDSS spectrum - providing the first NLTE study of any halo BHB star - and reconstruct its 3D trajectory in the Galactic potential. J1539+0239 turns out as the fastest halo star known to date, with a Galactic rest-frame velocity of 694$^{+300}_{-221}$ km/s (full uncertainty range from Monte Carlo error propagation) at its current position. The extreme kinematics of the star allows a significant lower limit to be put on the halo mass in order to keep it bound, of M_halo$\ge1.7_{-1.1}^{+2.3}\times10^{12}$ Msun. We conclude that the Xue et al. results tend to underestimate the true halo mass as their most likely mass value is consistent with our analysis only at a level of 4%. However, our result confirms other studies that make use of the full phase-space information.