Search

Both tin monosulfide (SnS) and tin disulfide (SnS2) are thermodynamically stable layered materials with potential for spin-valleytronic devices and photodetectors. Notably, monolayer SnS, owing to its low symmetry, exhibits interesting properties such as ferroelectricity, shift-current, and a persistent spin helix state in the monolayer limit. However, creating atomic-thickness crystals of SnS is challenging owing to the enhanced interlayer interactions caused by lone pair electrons, unlike to SnS2. Here, we demonstrate that p-type SnS can be selectively grown by simply varying the sulfur vapor concentration relative to tin using single-element precursors. We show that monolayer SnS crystals, up to several tens of micrometers in lateral scale, can be easily and safely obtained by high-temperature etching of bulk SnS in a pure nitrogen gas atmosphere. These findings pave the way for device applications based on high-quality tin sulfide.

Bi$_4$Br$_4$ is a prototypical quasi one-dimensional (1D) material in which covalently bonded bismuth bromide chains are arranged in parallel, side-by-side and layer-by-layer, with van der Waals (vdW) gaps in between. So far, two different structures have been reported for this compound, $\alpha$-Bi$_4$Br$_4$ and $\beta$-Bi$_4$Br$_4$ , in both of which neighboring chains are shifted by $\mathbf{b}/2$, i.e., half a unit cell vector in the plane, but which differ in their vertical stacking. While the different layer arrangements are known to result in distinct electronic properties, the effect of possible in-plane shifts between the atomic chains remains an open question. Here, using scanning tunneling microscopy and spectroscopy (STM/STS), we report a new Bi$_4$Br$_4$(001) structure, with a shift of $\mathbf{b}/3$ between neighboring chains in the plane and AB layer stacking. We determine shear strain to be the origin of this new structure, which can readily result in shifts of neighboring atomic chains because of the weak inter-chain bonding. For the observed $b/3$ structure, the (residual) atomic chain shift corresponds to an in-plane shear strain of $\gamma\approx7.5\%$. STS reveals a bulk insulating gap and metallic edge states at surface steps, indicating that the new structure is also a higher-order topological insulator, just like $\alpha$-Bi$_4$Br$_4$, in agreement with density functional theory (DFT) calculations., Comment: 9 pages, 5 figures, + Supplement

As one of the most prevalent diseases worldwide, plaque formation in human arteries, known as atherosclerosis, is the focus of many research efforts. Previously, molecular communication (MC) models have been proposed to capture and analyze the natural processes inside the human body and to support the development of diagnosis and treatment methods. In the future, synthetic MC networks are envisioned to span the human body as part of the Internet of Bio-Nano Things (IoBNT), turning blood vessels into physical communication channels. By observing and characterizing changes in these channels, MC networks could play an active role in detecting diseases like atherosclerosis. In this paper, building on previous preliminary work for simulating an MC scenario in a plaque-obstructed blood vessel, we evaluate different analytical models for non-Newtonian flow and derive associated channel impulse responses (CIRs). Additionally, we add the crucial factor of flow pulsatility to our simulation model and investigate the effect of the systole-diastole cycle on the received particles across the plaque channel. We observe a significant influence of the plaque on the channel in terms of the flow profile and CIR across different emission times in the cycle. These metrics could act as crucial indicators for early non-invasive plaque detection in advanced future MC methods., Comment: 6 pages, 6 figures

We present contemporaneous high-angular-resolution millimeter imaging and visible polarimetric imaging of the nearby asymptotic giant branch (AGB) star W Hya to better understand the dynamics and dust formation within a few stellar radii. The star W Hya was observed in two vibrationally excited H2O lines at 268 and 251 GHz with ALMA at a spatial resolution of 16 x 20 mas and at 748 and 820 nm at a resolution of 26 x 27 mas with the VLT/SPHERE-ZIMPOL. ALMA's high spatial resolution allowed us to image strong emission of the vibrationally excited H2O line at 268 GHz (v2 = 2, J_K_a,K_c = 6_5,2 - 7_4,3) over the stellar surface instead of absorption against the continuum, which is expected for thermal excitation. Strong, spotty emission was also detected along and just outside the stellar disk limb at an angular distance of ~40 mas (~1.9 stellar radii), extending to ~60 mas (~2.9 stellar radii). Another H2O line (v2 = 2, J_K_a,K_c = 9_2,8 - 8_3,5) at 251 GHz with a similar upper-level energy was tentatively identified, which shows absorption over the stellar surface. This suggests that the emission over the surface seen in the 268 GHzH2O line is suprathermal or even maser emission. The estimated gas temperature and H2O density are consistent with the radiatively pumped masers. The 268 GHz H2O line reveals global infall at up to ~15 km/s within 2--3 stellar radii, but outflows at up to ~8 km/s are also present. The polarized intensity maps obtained in the visible reveal clumpy dust clouds forming within ~40 mas (~1.9 stellar radii) with a particularly prominent cloud in the SW quadrant and a weaker cloud in the east. The 268 GHz H2O emission overlaps very well with the visible polarized intensity maps, which suggests that both the nonthermal and likely maser H2O emission and the dust originate from dense, cool pockets in the inhomogeneous atmosphere within ~2--3 stellar radii., Comment: 10 pages, 7 figures, published in Astronomy and Astrophysics

Owing to their rapid cooling rate and hence loss-limited propagation distance, cosmic-ray electrons and positrons (CRe) at very high energies probe local cosmic-ray accelerators and provide constraints on exotic production mechanisms such as annihilation of dark matter particles. We present a high-statistics measurement of the spectrum of CRe candidate events from 0.3 to 40 TeV with the High Energy Stereoscopic System (H.E.S.S.), covering two orders of magnitude in energy and reaching a proton rejection power of better than $10^{4}$. The measured spectrum is well described by a broken power law, with a break around 1 TeV, where the spectral index increases from $\Gamma_1 = 3.25$ $\pm$ 0.02 (stat) $\pm$ 0.2 (sys) to $\Gamma_2 = 4.49$ $\pm$ 0.04 (stat) $\pm$ 0.2 (sys). Apart from the break, the spectrum is featureless. The absence of distinct signatures at multi-TeV energies imposes constraints on the presence of nearby CRe accelerators and the local CRe propagation mechanisms., Comment: main paper: 8 pages, 4 figures, supplemental material: 12 pages, 14 figures, accepted for publication in Physical Review Letters https://journals.aps.org/prl/

Degenerate two-dimensional electron liquids are theoretically established to possess two vastly distinct collisional electron mean free paths, where even-parity deformations of the Fermi surface are hydrodynamic with a short collisional mean free path but odd-parity deformations remain near ballistic (known as the "tomographic" transport regime). Predicted signatures of this regime rely on the scaling of observables with temperature or device dimension, both of which are difficult to establish with certainty. Here, we consider magnetotransport in a minimal model of tomographic electrons and show that even a small magnetic field suppresses tomographic transport signatures and thus acts as a sensitive and unique probe of this regime. Fundamentally, the magnetic field breaks time-reversal invariance, which is a prerequisite for the odd-even parity effect in the collisional relaxation. We analyze in detail the scaling of the transverse conductivity, which has been linked to small-channel conductance of interaction-dominated electrons, and show that a tomographic scaling regime at intermediate wavenumbers is quickly suppressed with magnetic field to a hydrodynamic or collisionless form. We confirm that the suppression occurs at relatively small magnetic fields when the cyclotron radius is comparable to the ballistic mean free path of the dominant odd-parity mode. This occurs at a much smaller magnetic field than the magnetic field strength required to suppress hydrodynamic electron transport, which suggests an experimental protocol to extract the odd-parity mean free path., Comment: 17 pages, 7 figures

What mechanisms underlie linguistic generalization in large language models (LLMs)? This question has attracted considerable attention, with most studies analyzing the extent to which the language skills of LLMs resemble rules. As of yet, it is not known whether linguistic generalization in LLMs could equally well be explained as the result of analogical processes, which can be formalized as similarity operations on stored exemplars. A key shortcoming of prior research is its focus on linguistic phenomena with a high degree of regularity, for which rule-based and analogical approaches make the same predictions. Here, we instead examine derivational morphology, specifically English adjective nominalization, which displays notable variability. We introduce a new method for investigating linguistic generalization in LLMs: focusing on GPT-J, we fit cognitive models that instantiate rule-based and analogical learning to the LLM training data and compare their predictions on a set of nonce adjectives with those of the LLM, allowing us to draw direct conclusions regarding underlying mechanisms. As expected, rule-based and analogical models explain the predictions of GPT-J equally well for adjectives with regular nominalization patterns. However, for adjectives with variable nominalization patterns, the analogical model provides a much better match. Furthermore, GPT-J's behavior is sensitive to the individual word frequencies, even for regular forms, a behavior that is consistent with an analogical account of regular forms but not a rule-based one. These findings refute the hypothesis that GPT-J's linguistic generalization on adjective nominalization involves rules, suggesting similarity operations on stored exemplars as the underlying mechanism. Overall, our study suggests that analogical processes play a bigger role in the linguistic generalization of LLMs than previously thought.

The performance of embodied agents has been shown to improve by increasing model parameters, dataset size, and compute. This has been demonstrated in domains from robotics to video games, when generative learning objectives on offline datasets (pre-training) are used to model an agent's behavior (imitation learning) or their environment (world modeling). This paper characterizes the role of scale in these tasks more precisely. Going beyond the simple intuition that `bigger is better', we show that the same types of power laws found in language modeling (e.g. between loss and optimal model size), also arise in world modeling and imitation learning. However, the coefficients of these laws are heavily influenced by the tokenizer, task \& architecture -- this has important implications on the optimal sizing of models and data.

We propose MegaPortrait. It's an innovative system for creating personalized portrait images in computer vision. It has three modules: Identity Net, Shading Net, and Harmonization Net. Identity Net generates learned identity using a customized model fine-tuned with source images. Shading Net re-renders portraits using extracted representations. Harmonization Net fuses pasted faces and the reference image's body for coherent results. Our approach with off-the-shelf Controlnets is better than state-of-the-art AI portrait products in identity preservation and image fidelity. MegaPortrait has a simple but effective design and we compare it with other methods and products to show its superiority., Comment: Technical Report

We determine parameters of the renormalization group-consistent three-flavor color-superconducting Nambu-Jona-Lasinio (NJL) model that are suited to investigate possible compact-star configurations. Our goal is to provide quark-matter equations of state (EoS) that can be used for hadron-quark hybrid-star constructions. To that end, we mainly focus on the parameters of the quark-matter model. By varying the vector and diquark coupling constants, we analyze their impact on the EoS, the speed of sound, the maximum diquark gap, and the mass-radius relation. In almost all configurations, a stable color-flavor-locked (CFL) phase appears in the core of the maximum-mass configurations, typically spanning several kilometers in radius. In other cases, the star's two-flavor color-superconducting (2SC) branch of the EoS becomes unstable before reaching the CFL transition density. At neutron-star densities, the speed of sound squared reaches up to $c_s^2 \sim 0.6$ and the CFL gap up to $\Delta\sim250\,$MeV. We argue that adding a hadronic EoS at lower densities by performing a Maxwell construction does not increase the maximum mass substantially. Thus we use the $2.0 M_{\odot}$ constraint to constrain the NJL model parameters that are suited for the construction of hybrid-star EoS. We construct three examples of the hybrid-star model, demonstrating that there is room for different color-superconducting compositions. The hybrid EoSs obtained in this way can have no 2SC matter or different ratios of 2SC and CFL quark matter in the core. We show that early hadron-quark transitions are possible that can modify the tidal deformability at 1.4 $M_\odot$. We find that these EoSs are consistent with the imposed constraints from astrophysics and perturbative QCD. They allow for different hybrid-star scenarios with a hadronic EoS that is soft at low to intermediate densities ($\sim 1-3\, n_{\text{sat}}$)., Comment: 21 pages, 12 figures, 2 tables, comments are welcomed

Holes in planar germanium (Ge) heterostructures show promise for quantum applications, particularly in superconducting and spin qubits, due to strong spin-orbit interaction, low effective mass, and absence of valley degeneracies. However, charge traps cause issues such as gate hysteresis and charge noise. This study examines the effect of surface treatments on the accumulation behaviour and transport properties of Ge-based two dimensional hole gases (2DHGs). Oxygen plasma treatment reduces conduction in a setting without applied top-gate voltage and improves the mobility and lowers the percolation density, while hydrofluoric acid (HF) etching provides no benefit. The results suggest that interface traps from the partially oxidised silicon (Si) cap pin the Fermi level, and that oxygen plasma reduces the trap density by fully oxidising the Si cap. Therefore, optimising surface treatments is crucial for minimising the charge traps and thereby enhancing the device performance., Comment: 4 pages, 3 figures, additional supplementary information

Systems of correlated quantum matter can be a steep challenge to any would-be method of solution. Matrix-product state (MPS)-based methods can describe 1D systems quasiexactly, but often struggle to retain sufficient bipartite entanglement to accurately approximate 2D systems already. Conversely, Quantum Monte Carlo (QMC) approaches, based on sampling a probability distribution, can generally approximate 2D and 3D systems with an error that decays systematically with growing sampling size. However, QMC can suffer from the so-called sign problem, that makes the approach prohibitively costly for many systems of interest, such as repulsively interacting fermions away from commensurate densities and frustrated systems. In this article, we introduce a new hybrid approach, that combines auxiliary-field QMC (AFQMC) with MPS-based algorithms. This hybrid technique removes or reduces the sign problem (depending on the specific model) while also needing to retain much lower bipartite entanglement than brute-force application of a MPS-algorithm, without the use of uncontrolled approximations. We present two use-cases of the algorithm that would be challenging or impossible to address with any other approach, and quantify the extent of any remaining sign problem.

Understanding how neural networks transform input data across layers is fundamental to unraveling their learning and generalization capabilities. Although prior work has used insights from kernel methods to study neural networks, a global analysis of how the similarity between hidden representations evolves across layers remains underexplored. In this paper, we introduce a theoretical framework for the evolution of the kernel sequence, which measures the similarity between the hidden representation for two different inputs. Operating under the mean-field regime, we show that the kernel sequence evolves deterministically via a kernel map, which only depends on the activation function. By expanding activation using Hermite polynomials and using their algebraic properties, we derive an explicit form for kernel map and fully characterize its fixed points. Our analysis reveals that for nonlinear activations, the kernel sequence converges globally to a unique fixed point, which can correspond to orthogonal or similar representations depending on the activation and network architecture. We further extend our results to networks with residual connections and normalization layers, demonstrating similar convergence behaviors. This work provides new insights into the implicit biases of deep neural networks and how architectural choices influence the evolution of representations across layers.

The difficulty for solving ill-posed linear operator equations in Hilbert space is reflected by the strength of ill-posedness of the governing operator, and the inherent solution smoothness. In this study we focus on the ill-posedness of the operator, and we propose a partial ordering for the class of all bounded linear operators which lead to ill-posed operator equations. For compact linear operators, there is a simple characterization in terms of the decay rates of the singular values. In the context of the validity of the spectral theorem the partial ordering can also be understood. We highlight that range inclusions yield partial ordering, and we discuss cases when compositions of compact and non-compact operators occur. Several examples complement the theoretical results.

The alignment of permanent dipole moments and the resulting spontaneous orientation polarization (SOP) is commonly observed in evaporated neat films of polar organic molecules and leads to a so-called giant surface potential. In case of mixed films, often enhanced molecular orientation is observed, i.e.\ a higher degree of alignment, in comparison to neat layers, if it is diluted into a suitable (non-polar) host. So far, different possible influences on molecular orientation have been discussed, the most prominent probably being the so-called surface equilibration model. In this contribution, we discuss how surface equilibration can influence orientation in mixed layers, and which other intermolecular interactions have to be considered to explain the observed enhancement of SOP in mixed layers., Comment: 12 pages, 6 figures

This paper proposes a nonlinear magnetics model for Permanent Magnet Synchronous Machines (PMSMs) that accurately captures the effects of magnetic saturation in the machine iron and variations in rotor temperature on the permanent magnet excitation. The proposed model considers the permanent magnet as a current source rather than the more commonly used flux-linkage source. A comparison of the two modelling approaches is conducted using Finite Element Analysis (FEA) for different machine designs as well as experimental validation, where it is shown that the proposed model has substantially better accuracy. The proposed model decouples magnetic saturation and rotor temperature effects in the current/flux-linkage relationship, allowing for adaptive estimation of the PM excitation.

Language is not monolithic. While many benchmarks are used as proxies to systematically estimate Large Language Models' (LLM) performance in real-life tasks, they tend to ignore the nuances of within-language variation and thus fail to model the experience of speakers of minority dialects. Focusing on African American Vernacular English (AAVE), we present the first study on LLMs' fairness and robustness to a dialect in canonical reasoning tasks (algorithm, math, logic, and comprehensive reasoning). We hire AAVE speakers, including experts with computer science backgrounds, to rewrite seven popular benchmarks, such as HumanEval and GSM8K. The result of this effort is ReDial, a dialectal benchmark comprising $1.2K+$ parallel query pairs in Standardized English and AAVE. We use ReDial to evaluate state-of-the-art LLMs, including GPT-4o/4/3.5-turbo, LLaMA-3.1/3, Mistral, and Phi-3. We find that, compared to Standardized English, almost all of these widely used models show significant brittleness and unfairness to queries in AAVE. Furthermore, AAVE queries can degrade performance more substantially than misspelled texts in Standardized English, even when LLMs are more familiar with the AAVE queries. Finally, asking models to rephrase questions in Standardized English does not close the performance gap but generally introduces higher costs. Overall, our findings indicate that LLMs provide unfair service to dialect users in complex reasoning tasks. Code can be found at https://github.com/fangru-lin/redial_dialect_robustness_fairness.git.

Text generation, a key component in applications such as dialogue systems, relies on decoding algorithms that sample strings from a language model distribution. Traditional methods, such as top-$k$ and top-$\pi$, apply local normalisation to the model's output distribution, which can distort it. In this paper, we investigate the effect of this distortion by introducing globally-normalised versions of these decoding methods. Additionally, we propose an independent Metropolis-Hastings algorithm to approximate sampling from globally-normalised distributions without explicitly computing them. Our empirical analysis compares the performance of local and global normalisation across two decoding algorithms (top-$k$ and top-$\pi$) with various hyperparameters, using Pythia language models. Results show that, in most configurations, global decoding performs worse than the local decoding version of the same algorithms -- despite preserving the distribution's integrity. Our results suggest that distortion is an important feature of local decoding algorithms., Comment: Paper accepted in EMNLP 2024. Code is available in https://github.com/lowlypalace/global-decoding

An analog of nuclear magnetic resonance is realized in a microwave network with symplectic symmetry. The network consists of two identical subgraphs coupled by a pair of bonds with a length difference corresponding to a phase difference of $\pi$ for the waves traveling through the bonds. As a consequence all eigenvalues appear as Kramers doublets. Detuning the length difference from the $\pi$ condition Kramers degeneracy is lifted, which may be interpreted as a Zeeman splitting of a spin 1/2 in a magnetic field. The lengths of another pair of bonds are modulated periodically with frequencies of some 10 MHz by means of diodes, thus emulating a magnetic radiofrequency field. Features well-known from NMR such as the transition from the laboratory to the rotating frame, and Lorentzian shaped resonance curves can thus be realized., Comment: 5 pages, 4 figures. arXiv admin note: text overlap with arXiv:2410.07031

In a previous paper, we realized a microwave network with symplectic symmetry simulating a spin 1/2 (Rehemanjiang et al. [Phys. Rev. Lett. 117, 064101 (2016)]), following a suggestion by Joyner et al. [Europhys. Lett. 107, 50004(2014))]. The network consisted of two identical sub-units coupled by a pair of bonds with a length difference corresponding to a phase difference of $\pi$ for the waves traveling through the bonds. In such a symmetry each eigenvalue appears as a two-fold degenerate Kramers doublet. Distorting the symmetry the degeneracy is lifted which may be interpreted in terms of the Zeeman splitting of a spin 1/2 in an external magnetic field. In the present work, a microwave analog of a spin resonance is realized. To this end, two magnetic fields have to be emulated, a static and a radio-frequency one. The static one is realized by detuning the length difference from the $\pi$ condition by means of phase shifters, the radio-frequency field by modulating the length difference of another pair of bonds by means of diodes with frequencies up to 125 MHz. Features well-known from magnetic resonance such as the transition from the laboratory to the rotating frame, and Lorentzian shaped resonance curves can thus be realized., Comment: 13 pages, 8 figures

Evaluating diffusion-based image-editing models is a crucial task in the field of Generative AI. Specifically, it is imperative to assess their capacity to execute diverse editing tasks while preserving the image content and realism. While recent developments in generative models have opened up previously unheard-of possibilities for image editing, conducting a thorough evaluation of these models remains a challenging and open task. The absence of a standardized evaluation benchmark, primarily due to the inherent need for a post-edit reference image for evaluation, further complicates this issue. Currently, evaluations often rely on established models such as CLIP or require human intervention for a comprehensive understanding of the performance of these image editing models. Our benchmark, PixLens, provides a comprehensive evaluation of both edit quality and latent representation disentanglement, contributing to the advancement and refinement of existing methodologies in the field., Comment: 35 pages (17 main paper, 18 appendix), 22 figures

The precise and safe control of heavy material handling machines presents numerous challenges due to the hard-to-model hydraulically actuated joints and the need for collision-free trajectory planning with a free-swinging end-effector tool. In this work, we propose an RL-based controller that commands the cabin joint and the arm simultaneously. It is trained in a simulation combining data-driven modeling techniques with first-principles modeling. On the one hand, we employ a neural network model to capture the highly nonlinear dynamics of the upper carriage turn hydraulic motor, incorporating explicit pressure prediction to handle delays better. On the other hand, we model the arm as velocity-controllable and the free-swinging end-effector tool as a damped pendulum using first principles. This combined model enhances our simulation environment, enabling the training of RL controllers that can be directly transferred to the real machine. Designed to reach steady-state Cartesian targets, the RL controller learns to leverage the hydraulic dynamics to improve accuracy, maintain high speeds, and minimize end-effector tool oscillations. Our controller, tested on a mid-size prototype material handler, is more accurate than an inexperienced operator and causes fewer tool oscillations. It demonstrates competitive performance even compared to an experienced professional driver., Comment: Presented at IROS 2024, Abu Dhabi, as oral presentation

Golog is an expressive high-level agent language that includes nondeterministic operators which allow to leave some of the decisions to be made only at execution time. This so-called program realization is typically implemented by means of search, or in an incremental online fashion. In this paper, we consider the more realistic case where parts of the non-determinism are under the control of the environment. Program realization then becomes a synthesis problem, where a successful realization executes the program and satisfies the temporal goal for all possible environment actions. We consider Golog programs in combination with an expressive class of first-order action theories that allow for an unbounded number of objects and non-local effects, together with a temporal goal specified in a first-order extension of LTLf. We solve the synthesis problem by constructing a game arena that captures all possible executions of the program while tracking the satisfaction of the temporal goal and then solving the resulting two-player game. We evaluate the approach in two domains, showing the general feasibility of the approach.

2D materials research has reached significant scientific milestones, accompanied by a rapidly growing industrial sector in the two decades since the field's inception. Such rapid progress requires pushing past the boundary of what is technically and scientifically feasible and carries the risk of disseminating irreproducible research results. This Expert Recommendation addresses the need for enhanced reproducibility in 2D materials science and physics. Through a comprehensive examination of the factors that affect reproducibility the authors present a set of concrete guidelines designed to improve the reliability of research results. The introduction of a Standardised Template for Experimental Procedures (STEP) offers a novel approach to documenting experimental details that are crucial for replication and troubleshooting. We emphasise the importance of involving stakeholders from research, industry, publishing, funding agencies, and policymaking to foster a culture of transparency, reliability, and trust without blind angles and critical oversights. By addressing systemic issues that hinder reproducibility and presenting actionable steps for improvement, we aim to pave the way for more robust research practices in 2D materials science, contributing to the field's scientific maturation and the subsequent development of beneficial technologies., Comment: 24 pages, 3 textboxes, 1 figure, 2 tables

In this article we illustrate and draw connections between the geometry of zero sets of eigenfunctions, graph theory, vanishing order of eigenfunctions, and unique continuation. We identify the nodal set of an eigenfunction of the Laplacian (with smooth potential) on a compact, orientable Riemannian manifold as an \emph{imbedded metric graph} and then use tools from elementary graph theory in order to estimate the number of critical points in the nodal set of the $k$-th eigenfunction and the sum of vanishing orders at critical points in terms of $k$ and the genus of the manifold.

At a time when the influence of generative Artificial Intelligence on visual arts is a highly debated topic, we raise the attention towards a more subtle phenomenon: the algorithmic censorship of artistic nudity online. We analyze the performance of three "Not-Safe-For-Work'' image classifiers on artistic nudity, and empirically uncover the existence of a gender and a stylistic bias, as well as evident technical limitations, especially when only considering visual information. Hence, we propose a multi-modal zero-shot classification approach that improves artistic nudity classification. From our research, we draw several implications that we hope will inform future research on this topic., Comment: To be published at the AI4VA (AI for Visual Arts) Workshop and Challenges at ECCV 2024

Author profiling is the task of inferring characteristics about individuals by analyzing content they share. Supervised machine learning still dominates automatic systems that perform this task, despite the popularity of prompting large language models to address natural language understanding tasks. One reason is that the classification instances consist of large amounts of posts, potentially a whole user profile, which may exceed the input length of Transformers. Even if a model can use a large context window, the entirety of posts makes the application of API-accessed black box systems costly and slow, next to issues which come with such "needle-in-the-haystack" tasks. To mitigate this limitation, we propose a new method for author profiling which aims at distinguishing relevant from irrelevant content first, followed by the actual user profiling only with relevant data. To circumvent the need for relevance-annotated data, we optimize this relevance filter via reinforcement learning with a reward function that utilizes the zero-shot capabilities of large language models. We evaluate our method for Big Five personality trait prediction on two Twitter corpora. On publicly available real-world data with a skewed label distribution, our method shows similar efficacy to using all posts in a user profile, but with a substantially shorter context. An evaluation on a version of these data balanced with artificial posts shows that the filtering to relevant posts leads to a significantly improved accuracy of the predictions., Comment: preprint, under review, supplementary material will be made available upon acceptance of the paper

We present a local framework for investigating non-unitary evolution groups pertinent to effective field theories in general semi-classical spacetimes. Our approach is based on a rigorous local stability analysis of the algebra of observables and solely employs geometric concepts in the functional representation of quantum field theory. In this representation, it is possible to construct infinitely many self-adjoint extensions of the canonical momentum field at the kinematic level, and by the usual functional calculus arguments this holds for the Hamiltonian, as well. However, these self-adjoint domains have only the trivial wave functional in common with the solution space of the functional Schr\"odinger equation. This is related to the existence of boundaries in configuration field space which can be penetrated by the probability flux, causing probability to leak into regions in configuration field space that require a more fundamental description. As a consequence the evolution admits no unitary representation. Instead, in the absence of ghosts, the evolution is represented by contractive semi-groups in the semiclassical approximation. This allows to quantify the unitarity loss and, in turn, to assess the quality of the semi-classical approximation. We perform numerical experiments based on our formal investigations to determine regions in cosmological spacetimes where the semiclassical approximation breaks down for free quantum fields., Comment: 20 pages, 7 figures

Granville-Soundararajan, Harper-Nikeghbali-Radziwill, and Heap-Lindqvist independently established an asymptotic for the even natural moments of partial sums of random multiplicative functions defined over integers. Building on these works, we study the even natural moments of partial sums of Steinhaus random multiplicative functions defined over function fields. Using a combination of analytic arguments and combinatorial arguments, we obtain asymptotic expressions for all the even natural moments in the large field limit and large degree limit, as well as an exact expression for the fourth moment., Comment: 32 pages

P-value functions are modern statistical tools that unify effect estimation and hypothesis testing and can provide alternative point and interval estimates compared to standard meta-analysis methods, using any of the many p-value combination procedures available (Xie et al., 2011, JASA). We provide a systematic comparison of different combination procedures, both from a theoretical perspective and through simulation. We show that many prominent p-value combination methods (e.g. Fisher's method) are not invariant to the orientation of the underlying one-sided p-values. Only Edgington's method, a lesser-known combination method based on the sum of p-values, is orientation-invariant and provides confidence intervals not restricted to be symmetric around the point estimate. Adjustments for heterogeneity can also be made and results from a simulation study indicate that the approach can compete with more standard meta-analytic methods., Comment: with Appendix: 54 pages, 29 figures, 3 tables without Appendix: 31 pages, 8 figures, 3 tables

We demonstrate low energy, forming and compliance-free operation of a resistive memory obtained by the partial oxidation of a two-dimensional layered van-der-Waals semiconductor: hafnium disulfide (HfS$_2$). Semiconductor - oxide heterostructures are achieved by low temperature ($<300^{o}$C) thermal oxidation of HfS$_2$ in dry conditions, carefully controlling process parameters. The resulting HfO$_x$S$_y$/HfS$_2$ heterostructures are integrated between metal contacts, forming vertical crossbar devices. Forming-free, compliance-free resistive switching between non-volatile states is demonstrated by applying voltage pulses and measuring the current response in time. We show non-volatile memory operation with an R$_{ON}$/ R$_{OFF}$ of 102, programmable by 80ns WRITE and ERASE operations. Multiple stable resistance states are achieved by modulating pulse width and amplitude, down to 60ns, $<$ 20pJ operation. This demonstrates the capability of these devices for low - energy, fast-switching and multi-state programming. Resistance states were retained without fail at 150$^o$C over 10$^4$s, showcasing the potential of these devices for long retention times and resilience to ageing. Low-energy resistive switching measurements were repeated in vacuum (8.6 mbar) showing unchanged characteristics and no dependence of the device on surrounding oxygen or water vapour. Using a technology computer-aided design (TCAD) tool, we explore the role of the semiconductor layer in tuning the device conductance and driving gradual resistive switching in 2D HfO$_x$ - based devices.

The Nambu-Jona-Lasinio (NJL) model and specifically its extension to color superconductivity (CSC) is a popular effective model for investigating dense quark matter. However, the reliability of its results is challenged by cutoff artifacts, which emerge if temperature or chemical potential are of the order of the cutoff energy scales. In this work, we generalize an idea from [Braun et al. SciPost Phys., 6:056, 2019], which is based on the requirement of renormalization-group (RG) consistency and has successfully been applied to the two-flavor Quark-Meson-Diquark model, to the NJL model for electrically and color-neutral three-flavor color-superconducting quark matter. To this end, we analyze the medium divergences of the model and eliminate them by appropriate counterterms, introducing three different schemes. We show that the RG-consistent treatment removes the cutoff artifacts of the conventional regularization and enables the investigation of CSC matter at higher densities by the model. Our studies reveal the emergence of a so-called dSC phase within the melting pattern of the Color-Flavor Locked (CFL) phase at high chemical potentials, consistent with earlier Ginzburg-Landau analyses., Comment: 28 pages, 8 figures, 1 table

Compact metric spaces form an important class of metric spaces, but the category that they define lacks many important properties such as completeness and cocompleteness. In recent studies of "metric domain theory" and Stone-type dualities, the more general notion of a (separated) metric compact Hausdorff space emerged as a metric counterpart of Nachbin's compact ordered spaces. Roughly speaking, a metric compact Hausdorff space is a metric space equipped with a \emph{compatible} compact Hausdorff topology (which does not need to be the induced topology). These spaces maintain many important features of compact metric spaces, and, notably, the resulting category is much better behaved. Moreover, one can use inspiration from the theory of Nachbin's compact ordered spaces to solve problems for metric structures. In this paper we continue this line of research: in the category of separated metric compact Hausdorff spaces we characterise the regular monomorphisms as the embeddings and the epimorphisms as the surjective morphisms. Moreover, we show that epimorphisms out of an object $X$ can be encoded internally on $X$ by their kernel metrics, which are characterised as the continuous metrics below the metric on $X$; this gives a convenient way to represent quotient objects. Finally, as the main result, we prove that its dual category has an algebraic flavour: it is Barr-exact. While we show that it cannot be a variety of finitary algebras, it remains open whether it is an infinitary variety.

Motivated by the recent surge of interest in the geometry of hybrid spaces, we prove an Abel-Jacobi theorem for a metrized complex of Riemann surfaces, generalizing both the classical Abel-Jacobi theorem and its tropical analogue., Comment: 18 pages, 3 figures, comments very welcome!

Most two-dimensional (2D) materials experimentally studied so far have hexagons as their building blocks. Only a few exceptions, such as PdSe2, are lower in energy in pentagonal phases and exhibit pentagons as building blocks. While theory has predicted a large number of pentagonal 2D materials, many of them are metastable and their experimental realization is difficult. Here we report the successful synthesis of a metastable pentagonal 2D material, the monolayer pentagonal PdTe2, by symmetry-driven epitaxy. Scanning tunneling microscopy and complementary spectroscopy measurements are used to characterize the monolayer pentagonal PdTe2, which demonstrates well-ordered low-symmetry atomic arrangements and is stabilized by lattice matching with the underlying Pd(100) substrate. Theoretical calculations, along with angle-resolved photoemission spectroscopy, reveal monolayer pentagonal PdTe2 is a semiconductor with an indirect bandgap of 1.05 eV. Our work opens an avenue for the synthesis of pentagon-based 2D materials and gives opportunities to explore their applications such as multifunctional nanoelectronics.

It has recently been predicted that two-dimensional electron gases possess an anomalous ``tomographic'' transport regime outside of the traditional collisionless and hydrodynamic limits, but an experimental confirmation has been elusive so far. This anomalous regime is marked by the appearance of an odd-even effect in the quasiparticle lifetimes where deformations of the Fermi surface with odd-parity become long-lived in comparison to even-parity ones. In this work, we establish neutral atomic quantum gases as an alternative platform to reveal this new transport regime and demonstrate an odd-even effect in the normal phase of two-component Fermi gases. By diagonalizing the Fermi liquid collision integral, we identify odd-parity modes with anomalously long lifetimes below temperatures $T\leq 0.1 T_F$, which is within the reach of current cold atom experiments. In a marked difference from condensed matter setups, we show that the odd-even effect in neutral gases is widely tunable with interactions along the BCS-BEC crossover and suppressed on the BEC side where the Fermi surface is destroyed. We propose the damping rate of quadrupole oscillations as an experimental signature of the long-lived odd-parity modes. The damping rate is set by the shear viscosity, which for finite trap confinement is dominated by odd-parity modes and thus anomalous enhanced compared to the hydrodynamic limit. Furthermore, a full computation of the shear viscosity within Fermi liquid theory shows that the magnitude of the odd-even effect depends on the particle number and is particularly pronounced in mesoscopic Fermi gases. Our findings suggest that the hydrodynamic behavior of neutral degenerate quantum gases is much richer than previously thought and should include additional long-lived modes., Comment: 8 pages, 2 figures. Supplementary material: 4 pages, 3 figures

Covering ill-posed problems with compact and non-compact operators regarding the degree of ill-posedness is a never ending story written by many authors in the inverse problems literature. This paper tries to add a new narrative and some new facets with respect to this story under the auspices of the spectral theorem. The latter states that any self-adjoint and bounded operator is unitarily equivalent to a multiplication operator on some (semi-finite) measure space. We will exploit this fact and derive a distribution function from the corresponding multiplier, the growth behavior of which at zero allows us to characterize the degree of ill-posedness. We prove that this new concept coincides with the well-known one for compact operators (by means of their singular values), and illustrate the implications along examples including the Hausdorff moment operator and convolutions.

This paper proposes a new method to propagate uncertainties undergoing nonlinear dynamics using the Koopman Operator (KO). Probability density functions are propagated directly using the Koopman approximation of the solution flow of the system, where the dynamics have been projected on a well-defined set of basis functions. The prediction technique is derived following both the analytical (Galerkin) and numerical (EDMD) derivation of the KO, and a least square reduction algorithm assures the recursivity of the proposed methodology., Comment: 27th Conference of Information Fusion ID 14

We give a uniform vanishing criterion for products in bounded cohomology. This allows us to reprove and extend previous vanishing results for cup products and Massey triple products in the bounded cohomology of free groups and in the equivariant bounded cohomology of group actions on CAT(0) cube complexes., Comment: 28 pages, 3 figures

Continual learning is the problem of integrating new information in a model while retaining the knowledge acquired in the past. Despite the tangible improvements achieved in recent years, the problem of continual learning is still an open one. A better understanding of the mechanisms behind the successes and failures of existing continual learning algorithms can unlock the development of new successful strategies. In this work, we view continual learning from the perspective of the multi-task loss approximation, and we compare two alternative strategies, namely local and global approximations. We classify existing continual learning algorithms based on the approximation used, and we assess the practical effects of this distinction in common continual learning settings.Additionally, we study optimal continual learning objectives in the case of local polynomial approximations and we provide examples of existing algorithms implementing the optimal objectives, Comment: (10 pages, Will appear in the proceedings of CoLLAs 2024)

The Tarantula Nebula in the Large Magellanic Cloud is known for its high star formation activity. At its center lies the young massive star cluster R136, providing a significant amount of the energy that makes the nebula shine so brightly at many wavelengths. Recently, young massive star clusters have been suggested to also efficiently produce high-energy cosmic rays, potentially beyond PeV energies. Here, we report the detection of very-high-energy $\gamma$-ray emission from the direction of R136 with the High Energy Stereoscopic System, achieved through a multicomponent, likelihood-based modeling of the data. This supports the hypothesis that R136 is indeed a very powerful cosmic-ray accelerator. Moreover, from the same analysis, we provide an updated measurement of the $\gamma$-ray emission from 30 Dor C, the only superbubble detected at TeV energies presently. The $\gamma$-ray luminosity above $0.5\,\mathrm{TeV}$ of both sources is $(2-3)\times 10^{35}\,\mathrm{erg}\,\mathrm{s}^{-1}$. This exceeds by more than a factor of 2 the luminosity of HESS J1646$-$458, which is associated with the most massive young star cluster in the Milky Way, Westerlund 1. Furthermore, the $\gamma$-ray emission from each source is extended with a significance of $>3\sigma$ and a Gaussian width of about $30\,\mathrm{pc}$. For 30 Dor C, a connection between the $\gamma$-ray emission and the nonthermal X-ray emission appears likely. Different interpretations of the $\gamma$-ray signal from R136 are discussed., Comment: 10+11 pages, 4+6 figures. Corresponding authors: L. Mohrmann, N. Komin

Planar semiconductor heterostructures offer versatile device designs and are promising candidates for scalable quantum computing. Notably, heterostructures based on strained germanium have been extensively studied in recent years, with emphasis on their strong and tunable spin-orbit interaction, low effective mass, and high hole mobility. However, planar systems are still limited by the fact that the shape of the confinement potential is directly related to the density. In this work, we present the successful implementation of a backgate for a planar germanium heterostructure. The backgate, in combination with a topgate, enables independent control over the density and the electric field, which determines important state properties such as the effective mass, the $g$-factor and the quantum lifetime. This unparalleled degree of control paves the way towards engineering qubit properties and facilitates the targetted tuning of bilayer quantum wells, which promise denser qubit packing., Comment: 14 pages plus 2 pages supplementary information, 5 figures in main text

We analyze the Kitaev model on the $\{9,3\}$ hyperbolic lattice. The $\{9,3\}$ is formed by a regular 3-coordinated tiling of nonagons, where the 3-color coding of bonds according to the inequivalent Kitaev Ising spin couplings yields the natural generalization of the original Kitaev model for Euclidean regular honeycomb tiling. Upon investigation of the bulk spectrum for large finite size droplets, we identify a gapless chiral $\mathbb{Z}_2$ spin liquid state featuring spontaneous time reversal symmetry breaking. Due to its non-commutative translation group structure, such type of hyperbolic spin liquid is conjectured to feature chiral quasiparticles with a potentially non-Abelian Bloch profile.

The upcoming 6G mobile communication standard will offer a revolutionary new feature: Integrated sensing and communication (ISAC) reuses mobile communication signals to realize multi-static radar for various applications including localization. Consequently, applied ISAC propagation research necessitates to evolve from classical monostatic radar cross section (RCS) measurement of static targets on to bistatic radar reflectivity characterization of dynamic objects. Here, we introduce our "Bistatic Radar" (BIRA) and antenna measurement facility for bistatic spherical positioning with sub-millimeter accuracy on a diameter of up to 7 m and with almost continuous frequency coverage from 0.7 up to 260 GHz. Currently, BIRA is the only bistatic measurement facility capable of unrestricted ISAC research: In addition to vector network analysis, BIRA employs advanced wideband transceiver technology with an instantaneous bandwidth of up to 4 GHz. These transceivers grant BIRA the unique ability to characterize dynamic targets in both Doppler and range, while also significantly accelerating RCS measurements of static objects., Comment: 9 pages, 8 figures

In multilingual settings, non-Latin scripts and low-resource languages are usually disadvantaged in terms of language models' utility, efficiency, and cost. Specifically, previous studies have reported multiple modeling biases that the current tokenization algorithms introduce to non-Latin script languages, the main one being over-segmentation. In this work, we propose MAGNET; multilingual adaptive gradient-based tokenization to reduce over-segmentation via adaptive gradient-based subword tokenization. MAGNET learns to predict segment boundaries between byte tokens in a sequence via sub-modules within the model, which act as internal boundary predictors (tokenizers). Previous gradient-based tokenization methods aimed for uniform compression across sequences by integrating a single boundary predictor during training and optimizing it end-to-end through stochastic reparameterization alongside the next token prediction objective. However, this approach still results in over-segmentation for non-Latin script languages in multilingual settings. In contrast, MAGNET offers a customizable architecture where byte-level sequences are routed through language-script-specific predictors, each optimized for its respective language script. This modularity enforces equitable segmentation granularity across different language scripts compared to previous methods. Through extensive experiments, we demonstrate that in addition to reducing segmentation disparities, MAGNET also enables faster language modelling and improves downstream utility.