Search

Inferring sky surface brightness distributions from noisy interferometric data in a principled statistical framework has been a key challenge in radio astronomy. In this work, we introduce Imaging for Radio Interferometry with Score-based models (IRIS). We use score-based models trained on optical images of galaxies as an expressive prior in combination with a Gaussian likelihood in the uv-space to infer images of protoplanetary disks from visibility data of the DSHARP survey conducted by ALMA. We demonstrate the advantages of this framework compared with traditional radio interferometry imaging algorithms, showing that it produces plausible posterior samples despite the use of a misspecified galaxy prior. Through coverage testing on simulations, we empirically evaluate the accuracy of this approach to generate calibrated posterior samples., Comment: 17 pages, 8 figures, submitted to the Astrophysical Journal

Creativity relates to the ability to generate novel and effective ideas in the areas of interest. How are such creative ideas generated? One possible mechanism that supports creative ideation and is gaining increased empirical attention is by asking questions. Question asking is a likely cognitive mechanism that allows defining problems, facilitating creative problem solving. However, much is unknown about the exact role of questions in creativity. This work presents an attempt to apply text mining methods to measure the cognitive potential of questions, taking into account, among others, (a) question type, (b) question complexity, and (c) the content of the answer. This contribution summarizes the history of question mining as a part of creativity research, along with the natural language processing methods deemed useful or helpful in the study. In addition, a novel approach is proposed, implemented, and applied to five datasets. The experimental results obtained are comprehensively analyzed, suggesting that natural language processing has a role to play in creative research., Comment: 24 pages, 15 figures; accepted to International Conference on Big Data Analytics in Astronomy, Science and Engineering 2024, Aizu, Japan

Although generative models hold promise for discovering molecules with optimized desired properties, they often fail to suggest synthesizable molecules that improve upon the known molecules seen in training. We find that a key limitation is not in the molecule generation process itself, but in the poor generalization capabilities of molecular property predictors. We tackle this challenge by creating an active-learning, closed-loop molecule generation pipeline, whereby molecular generative models are iteratively refined on feedback from quantum chemical simulations to improve generalization to new chemical space. Compared against other generative model approaches, only our active learning approach generates molecules with properties that extrapolate beyond the training data (reaching up to 0.44 standard deviations beyond the training data range) and out-of-distribution molecule classification accuracy is improved by 79%. By conditioning molecular generation on thermodynamic stability data from the active-learning loop, the proportion of stable molecules generated is 3.5x higher than the next-best model.

Clouds play a key role in Earth's radiation balance with complex effects that introduce large uncertainties into climate models. Real-time 3D cloud data is essential for improving climate predictions. This study leverages geostationary imagery from MSG/SEVIRI and radar reflectivity measurements of cloud profiles from CloudSat/CPR to reconstruct 3D cloud structures. We first apply self-supervised learning (SSL) methods-Masked Autoencoders (MAE) and geospatially-aware SatMAE on unlabelled MSG images, and then fine-tune our models on matched image-profile pairs. Our approach outperforms state-of-the-art methods like U-Nets, and our geospatial encoding further improves prediction results, demonstrating the potential of SSL for cloud reconstruction.

Simplicial complexes arising from real-world settings may not be directly observable. Hence, for an unknown simplicial complex in Euclidean space, we want to efficiently reconstruct it by querying local structure. In particular, we are interested in queries for the indegree of a simplex $\sigma$ in some direction: the number of cofacets of $\sigma$ contained in some halfspace "below" $\sigma$. Fasy et al. proposed a method that, given the vertex set of a simplicial complex, uses indegree queries to reconstruct the set of edges. In particular, they use a sweep algorithm through the vertex set, identifying edges adjacent to and above each vertex in the sweeping order. The algorithm relies on a natural but crucial property of the sweeping order: at a given vertex $v$, all edges adjacent to $v$ contained in the halfspace below $v$ have another endpoint that appeared earlier in the order. The edge reconstruction algorithm does not immediately extend to higher-dimensional simplex reconstruction. In particular, it is not possible to sweep through a set of $i$-simplices in a fixed direction and maintain that all $(i+1)$-cofacets of a given simplex $\sigma$ that come below $\sigma$ are known. We circumvent this by defining a sweeping order on a set of $i$-simplices, that additionally pairs each $i$-simplex $\sigma$ with a direction perpendicular to $\sigma$. Analogous to Fasy et al., our order has the crucial property that, at any $i$-simplex $\sigma$ paired with direction $s$, each $(i+1)$-dimensional coface of $\sigma$ that lies in the halfspace below $\sigma$ with respect to the direction $s$ has an $i$-dimensional face that appeared earlier in the order. We show how to compute such an order and use it to extend the edge reconstruction algorithm of Fasy et al. to simplicial complex reconstruction. Our algorithm can reconstruct arbitrary embedded simplicial complexes.

The world of nanoscales in fluidics is the frontier where the continuum of fluid mechanics meets the atomic, and even quantum, nature of matter. While water dynamics remains largely classical under extreme confinement, several experiments have recently reported coupling between water transport and the electronic degrees of freedom of the confining materials. This avenue prompts us to reconsider nanoscale hydrodynamic flows under the perspective of interacting excitations, akin to condensed matter frameworks. Here we show, using a combination of many-body theory and molecular simulations, that the flow of a liquid can induce the flow of another liquid behind a separating wall, at odds with the prediction of continuum hydrodynamics. We further show that the range of this 'flow tunnelling' can be tuned through the solid's electronic excitations, with a maximum occurring when these are at resonance with the liquid's charge density fluctuations. Flow tunnelling is expected to play a role in global transport across nanoscale fluidic networks, such as lamellar graphene oxide or MXene membranes. It further suggests exploiting the electronic properties of the confining walls for manipulating liquids via their dielectric spectra, beyond the nature and characteristics of individual molecules.

Reconstructing the intricate local morphology of neurons and their long-range projecting axons can address many connectivity related questions in neuroscience. The main bottleneck in connectomics pipelines is correcting topological errors, as multiple entangled neuronal arbors is a challenging instance segmentation problem. More broadly, segmentation of curvilinear, filamentous structures continues to pose significant challenges. To address this problem, we extend the notion of simple points from digital topology to connected sets of voxels (i.e. supervoxels) and propose a topology-aware neural network segmentation method with minimal computational overhead. We demonstrate its effectiveness on a new public dataset of 3-d light microscopy images of mouse brains, along with the benchmark datasets DRIVE, ISBI12, and CrackTree.

Barycentric coordinates provide solutions to the problem of expressing an element of a compact convex set as a convex combination of a finite number of extreme points of the set. They have been studied widely within the geometric literature, typically in response to the demands of interpolation, numerical analysis and computer graphics. In this note we bring an algebraic perspective to the problem, based on barycentric algebras. We focus on the discussion of relations between different subclasses of partitions of unity, one arising in the context of barycentric coordinates, based on the tautological map introduced by Guessab.

Consider $d+2$ i.i.d. random points $X_1,\ldots, X_{d+2}$ in $\mathbb R^d$. In this short note we compute the probability that their convex hull is a simplex in the following three cases: (i) the distribution of $X_1$ is multivariate standard normal; (ii) the density of $X_1$ is proportional to $(1-\|x\|^2)^{\beta}$ on the unit ball (beta distribution); (iii) the density of $X_1$ is proportional to $(1+\|x\|^2)^{-\beta}$ (beta prime distribution). In the normal case, the above probability is twice the sum of solid angles of a regular $(d+1)$-dimensional simplex., Comment: 12 pages

In this article, we provide a stochastic sampling algorithm with polynomial complexity in fixed dimension that leverages the recent advances on diffusion models where it is shown that under mild conditions, sampling can be achieved via an accurate estimation of intermediate scores across the marginals $(p_t)_{t\ge 0}$ of the standard Ornstein-Uhlenbeck process started at $\mu$, the density we wish to sample from. The heart of our method consists into approaching these scores via a computationally cheap estimator and relating the variance of this estimator to the smoothness properties of the forward process. Under the assumption that the density to sample from is $L$-log-smooth and that the forward process is semi-log-concave: $-\nabla^2 \log(p_t) \succeq -\beta I_d$ for some $\beta \geq 0$, we prove that our algorithm achieves an expected $\epsilon$ error in $\text{KL}$ divergence in $O(d^7(L+\beta)^2L^{d+2}\epsilon^{-2(d+3)}(d+m_2(\mu))^{2(d+1)})$ time with $m_2(\mu)$ the second order moment of $\mu$. In particular, our result allows to fully transfer the problem of sampling from a log-smooth distribution into a regularity estimate problem. As an application, we derive an exponential complexity improvement for the problem of sampling from an $L$-log-smooth distribution that is $\alpha$-strongly log-concave outside some ball of radius $R$: after proving that such distributions verify the semi-log-concavity assumption, a result which might be of independent interest, we recover a $poly(R, L, \alpha^{-1}, \epsilon^{-1})$ complexity in fixed dimension which exponentially improves upon the previously known $poly(e^{LR^2}, L,\alpha^{-1}, \log(\epsilon^{-1}))$ complexity in the low precision regime.

Here, we describe one of the first Web-scale hybrid Knowledge Graph (KG)-Large Language Model (LLM), populated with the latest peer-reviewed medical knowledge on colorectal Cancer. It is currently being evaluated to assist with both medical research and clinical information retrieval tasks at Moffitt Cancer Center, which is one of the top Cancer centers in the U.S. and in the world. Our hybrid is remarkable as it serves the user needs better than just an LLM, KG or a search-engine in isolation. LLMs as is are known to exhibit hallucinations and catastrophic forgetting as well as are trained on outdated corpora. The state of the art KGs, such as PrimeKG, cBioPortal, ChEMBL, NCBI, and other require manual curation, hence are quickly getting stale. CancerKG is unsupervised and is capable of automatically ingesting and organizing the latest medical findings. To alleviate the LLMs shortcomings, the verified KG serves as a Retrieval Augmented Generation (RAG) guardrail. CancerKG exhibits 5 different advanced user interfaces, each tailored to serve different data modalities better and more convenient for the user.

This work describes our group's submission to the PROCESS Challenge 2024, with the goal of assessing cognitive decline through spontaneous speech, using three guided clinical tasks. This joint effort followed a holistic approach, encompassing both knowledge-based acoustic and text-based feature sets, as well as LLM-based macrolinguistic descriptors, pause-based acoustic biomarkers, and multiple neural representations (e.g., LongFormer, ECAPA-TDNN, and Trillson embeddings). Combining these feature sets with different classifiers resulted in a large pool of models, from which we selected those that provided the best balance between train, development, and individual class performance. Our results show that our best performing systems correspond to combinations of models that are complementary to each other, relying on acoustic and textual information from all three clinical tasks.

Wireless radio frequency coils provide a promising solution for clinical MR applications due to several benefits, such as cable-free connection and compatibility with MR platforms of different vendors. Namely, for the purpose of clinical high-field human breast imaging several wireless transceiver coils are known to the date, those operational principle is based on inductive coupling with a body coil. These coils are commonly consist of a several volume resonators to perform bilateral breast imaging. Due to the electrically close location of volume resonators, strong inductive coupling is observed, resulting in the occurrence of hybrid modes. In principle, MR imaging using one of the hybrid modes is possible provided by the homogeneity of a B+ distribution. However, the question of influence of volume resonators coupling on wireless coil transmit efficiency and receive sensitivity was not previously studied. By this work, we performed study to understand this issue. The first wireless coil with decoupled resonators is developed, evaluated numerically and experimentally including in vivo study on healthy volunteers. According to the obtained results, transmit efficiency and receive sensitivity of a pair of decoupled Helmholtz resonators is at least 24% higher than for a pair of coupled resonators., Comment: 12 pages, 10 figures, 2 tables

We show the existence of traveling front solutions in a diffusive classical SIS epidemic model and the SIS model with a saturating incidence in the size of the susceptible population. We investigate the situation where both susceptible and infected populations move around at a comparable rates, but small compared to the spatial scale. In this case, we show that traveling front solutions exist for each fixed positive speed. In the regime where the infected population diffuses slower than the susceptible population, we show the existence of traveling wave solutions for each fixed positive speed and describe their structure and dependence on the wave speed which as it is varied from 0 to infinity. In the regime where the infected population diffuses faster than the susceptible population, we derive a bound for the speeds of the fronts in this regime in which the infection propagates as a front. Moreover, for the classical SIS model we show that there is a case when the spread of the disease is governed by the Burgers-FKPP equation., Comment: 17 pages

The landscape of digital servitization in the manufacturing sector is evolving, marked by a strategic shift from traditional product-centric to platform business models (BMs). Manufacturing firms often employ a blend of approaches to develop business-to-business (B2B) platforms, leading to significant reconfigurations in their BMs. However, they frequently encounter failures in their B2B platform development initiatives, leading them to abandon initial efforts and pivot to alternative platform strategies. Therefore, this study, through an in-depth case study of a manufacturer in the energy sector, articulates a three-phase pivoting framework for B2B platform BMs, including platform development and platform strategy. Initially, the manufacturer focused on asset-based product sales supplemented by asset maintenance services and followed an emergent platformization strategy characterized by the rise of multiple, independent B2B platforms catering to diverse functions. Next, focusing on the imposed customer journey strategy, the firm shifted towards a strategic multi-platform integration into an all-encompassing platform supported by artificial intelligence (AI), signaling a maturation of the platform BM to combine a wide range of services into an energy-performance-based contract. Finally, the last step of the firm's platform BM evolution consisted of a deliberate platform strategy open to external stakeholders and enveloping its data-driven offerings within a broader platform ecosystem. This article advances B2B platform BMs and digital servitization literature, highlighting the efficacy of a progressive approach and strategic pivoting.

Layer potentials represent solutions to partial differential equations in an integral equation formulation. When numerically evaluating layer potentials at evaluation points close to the domain boundary, specialized quadrature techniques are required for accuracy because of rapid variations in the integrand. To efficiently achieve a specified error tolerance, we introduce an adaptive quadrature method with automatic parameter adjustment for axisymmetric surfaces, facilitated by error estimation. Notably, while each surface must be axisymmetric, the integrand itself need not be, allowing for applications with complex geometries featuring multiple axisymmetric bodies. The proposed quadrature method utilizes so-called interpolatory semi-analytical quadrature in conjunction with a singularity swap technique in the azimuthal angle. In the polar angle, such a technique is used as needed, depending on the integral kernel, combined with an adaptive subdivision of the integration interval. The method is tied to a regular quadrature method that employs a trapezoidal rule in the azimuthal angle and a Gauss-Legendre quadrature rule in the polar angle, which will be used whenever deemed sufficiently accurate, as determined by a quadrature error estimate [C. Sorgentone and A.-K. Tornberg, Advances in Computational Mathematics, 49 (2023), p. 87]. Error estimates for both numerical integration and interpolation are derived using complex analysis, and are used to determine the adaptive panel subdivision given the evaluation point and desired accuracy. Numerical examples are presented to demonstrate the method's efficacy., Comment: 36 pages, 11 figures

Geological materials are often seen as the antithesis of soft; rocks are hard. However, during the formation of minerals and rocks, all the systems we shall discuss, indeed geological materials in general, pass through a stage where they are soft. This occurs either because they are at a high temperature -- igneous or metamorphic rock -- or because they are at a lower temperature but in the presence of water -- sedimentary rock. For this reason it is useful to introduce soft-matter concepts into the geological domain. There is a universality in the diverse instances of geological patterns that may be appreciated by looking at the common aspect in their formation of having passed through a stage as soft matter.

Let $f \colon \Omega \to \Omega' $ be a Sobolev mapping of finite distortion between planar domains $\Omega $ and $\Omega'$, satisfying the $(INV)$ condition and coinciding with a homeomorphism near $\partial\Omega $. We show that $f$ admits a generalized inverse mapping $h \colon \Omega' \to \Omega$, which is also a Sobolev mapping of finite distortion and satisfies the $(INV)$ condition. We also establish a higher-dimensional analogue of this result: if a mapping $f \colon \Omega \to \Omega' $ of finite distortion is in the Sobolev class $W^{1,p}(\Omega, \mathbb{R}^n)$ with $p > n-1$ and satisfies the $(INV)$ condition, then $f$ has an inverse in $W^{1,1}(\Omega', \mathbb{R}^n)$ that is also of finite distortion. Furthermore, we characterize Sobolev mappings satisfying $(INV)$ whose generalized inverses have finite $n$-harmonic energy.

In this paper we study Poisson processes of so-called "fat" cylinders in hyperbolic space. As our main result we show that this model undergoes a percolation phase transition. We prove this by establishing a novel link between the fat Poisson cylinder process and semi-scale invariant random fractal models on the unit sphere and in $\mathbb{R}^d.$ As a secondary result, we show that the semi-scale invariant fractal ball model in $\mathbb{R}^d$ has a non-empty sheet phase.

The tris(2,4,6-trichlorophenyl)methyl radical (TTM) has inspired the synthesis of several luminescent diradicals and diradicaloids, providing an extraordinary opportunity to control the nature of the low-lying excited states by fine-tuning the diradical character. However, the photophysical properties of TTM-derived diradicals remain not fully understood yet. Here we present a combined theoretical and experimental investigation on TTM-derived diradicals to elucidate the origin of their luminescence. The theoretical analysis focuses on a series of symmetric TTM-derived diradicals with singlet ground state, featuring radical moieties linked by pi-conjugated spacers of different length. The nature of the lowest excited electronic states that control their photophysical behaviour is discussed in detail. The study is complemented by a complete spectroscopic characterization of the TTM-TTM diradical, synthesized using a novel, simpler and more efficient procedure exploiting the unique reactivity of TTM. The lowest excited states of the diradicals differ qualitatively from those of TTM: two novel low-lying states emerge in the diradical, due to charge resonance (CR) between the two radical units. The lowest CR state is a dark state for symmetric diradicals. The CR nature explains the blue-shifted emission observed by increasing the distance between the radical centres as seen in TTM-ph-TTM. This insight suggests different design strategies to improve the luminescence properties of TTM-derived diradicals, Comment: 11 pages, 5 figures

The surge in online content has created an urgent demand for robust detection systems, especially in non-English contexts where current tools demonstrate significant limitations. We present forePLay, a novel Polish language dataset for erotic content detection, featuring over 24k annotated sentences with a multidimensional taxonomy encompassing ambiguity, violence, and social unacceptability dimensions. Our comprehensive evaluation demonstrates that specialized Polish language models achieve superior performance compared to multilingual alternatives, with transformer-based architectures showing particular strength in handling imbalanced categories. The dataset and accompanying analysis establish essential frameworks for developing linguistically-aware content moderation systems, while highlighting critical considerations for extending such capabilities to morphologically complex languages., Comment: The forePLay dataset and associated resources will be made publicly available for research purposes upon publication, in accordance with data sharing regulations

We introduce for the first time a neural-certificate framework for continuous-time stochastic dynamical systems. Autonomous learning systems in the physical world demand continuous-time reasoning, yet existing learnable certificates for probabilistic verification assume discretization of the time continuum. Inspired by the success of training neural Lyapunov certificates for deterministic continuous-time systems and neural supermartingale certificates for stochastic discrete-time systems, we propose a framework that bridges the gap between continuous-time and probabilistic neural certification for dynamical systems under complex requirements. Our method combines machine learning and symbolic reasoning to produce formally certified bounds on the probabilities that a nonlinear system satisfies specifications of reachability, avoidance, and persistence. We present both the theoretical justification and the algorithmic implementation of our framework and showcase its efficacy on popular benchmarks.

Fundamental research on two-dimensional (2D) magnetic systems based on van der Waals materials has been gaining traction rapidly since their recent discovery. With the increase of recent knowledge, it has become clear that such materials have also a strong potential for applications in devices that combine magnetism with electronics, optics, and nanomechanics. Nonetheless, many challenges still lay ahead. Several fundamental aspects of 2D magnetic materials are still unknown or poorly understood, such as their often-complicated electronic structure, optical properties, and magnetization dynamics, and their magnon spectrum. To elucidate their properties and facilitate integration in devices, advanced characterization techniques and theoretical frameworks need to be developed or adapted. Moreover, developing synthesis methods which increase critical temperatures and achieve large-scale, high-quality homogeneous thin films is crucial before these materials can be used for real-world applications. Therefore, the field of 2D magnetic materials provides many challenges and opportunities for the discovery and exploration of new phenomena, as well as the development of new applications. This Roadmap presents the background, challenges, and potential research directions for various relevant topics in the field on the fundamentals, synthesis, characterization, and applications. We hope that this work can provide a strong starting point for young researchers in the field and provide a general overview of the key challenges for more experienced researchers., Comment: 87 pages, 24 figures. Roadmap based on the discussions during the workshop on Quantum Magnetic Materials hosted by the Lorentz Centre in Leiden, the Netherlands (Oct. 2023)

General matrix multiplication (GEMM) is a ubiquitous computing kernel/algorithm for data processing in diverse applications, including artificial intelligence (AI) and deep learning (DL). Recent shift towards edge computing has inspired GEMM architectures based on unary computing, which are predominantly stochastic and rate-coded systems. This paper proposes a novel GEMM architecture based on temporal-coding, called tuGEMM, that performs exact computation. We introduce two variants of tuGEMM, serial and parallel, with distinct area/power-latency trade-offs. Post-synthesis Power-Performance-Area (PPA) in 45 nm CMOS are reported for 2-bit, 4-bit, and 8-bit computations. The designs illustrate significant advantages in area-power efficiency over state-of-the-art stochastic unary systems especially at low precisions, e.g. incurring just 0.03 mm^2 and 9 mW for 4 bits, and 0.01 mm^2 and 4 mW for 2 bits. This makes tuGEMM ideal for power constrained mobile and edge devices performing always-on real-time sensory processing., Comment: Published in 2023 IEEE International Symposium on Circuits and Systems (ISCAS), Monterey, CA, USA, 2023

General Matrix Multiplication (GEMM) is a ubiquitous compute kernel in deep learning (DL). To support energy-efficient edge-native processing, new GEMM hardware units have been proposed that operate on unary encoded bitstreams using much simpler hardware. Most unary approaches thus far focus on rate-based unary encoding of values and perform stochastic approximate computation. This work presents tubGEMM, a novel matrix-multiply unit design that employs hybrid temporal-unary and binary (tub) encoding and performs exact (not approximate) GEMM. It intrinsically exploits dynamic value sparsity to improve energy efficiency. Compared to the current best unary design uGEMM, tubGEMM significantly reduces area, power, and energy by 89\%, 87\%, and 50\%, respectively. A tubGEMM design performing 128x128 matrix multiply on 8-bit integers, in commercial TSMC N5 (5nm) process node, consumes just 0.22 mm^2 die area, 417.72 mW power, and 8.86 uJ energy, assuming no sparsity. Typical sparsity in DL workloads (MobileNetv2, ResNet-50) reduces energy by more than 3x, and lowering precision to 4 and 2 bits further reduces it by 24x and 104x respectively., Comment: Published in 2023 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)

We provide a review and a comparison of methods for differential network estimation in Gaussian graphical models with focus on structure learning. We consider the case of two datasets from distributions associated with two graphical models. In our simulations, we use five different methods to estimate differential networks. We vary graph structure and sparsity to explore their influence on performance in terms of power and false discovery rate. We demonstrate empirically that presence of hubs proves to be a challenge for all the methods, as well as increased density. We suggest local and global properties that are associated with this challenge. Direct estimation with lasso penalized D-trace loss is shown to perform the best across all combinations of network structure and sparsity., Comment: 20 pages, 5 figures

This paper introduces mixed-integer optimization methods to solve regression problems that incorporate fairness metrics. We propose an exact formulation for training fair regression models. To tackle this computationally hard problem, we study the polynomially-solvable single-factor and single-observation subproblems as building blocks and derive their closed convex hull descriptions. Strong formulations obtained for the general fair regression problem in this manner are utilized to solve the problem with a branch-and-bound algorithm exactly or as a relaxation to produce fair and accurate models rapidly. Moreover, to handle large-scale instances, we develop a coordinate descent algorithm motivated by the convex-hull representation of the single-factor fair regression problem to improve a given solution efficiently. Numerical experiments conducted on fair least squares and fair logistic regression problems show competitive statistical performance with state-of-the-art methods while significantly reducing training times.

Gaussian processes are now commonly used in dimensionality reduction approaches tailored to neuroscience, especially to describe changes in high-dimensional neural activity over time. As recording capabilities expand to include neuronal populations across multiple brain areas, cortical layers, and cell types, interest in extending Gaussian process factor models to characterize multi-population interactions has grown. However, the cubic runtime scaling of current methods with the length of experimental trials and the number of recorded populations (groups) precludes their application to large-scale multi-population recordings. Here, we improve this scaling from cubic to linear in both trial length and group number. We present two approximate approaches to fitting multi-group Gaussian process factor models based on (1) inducing variables and (2) the frequency domain. Empirically, both methods achieved orders of magnitude speed-up with minimal impact on statistical performance, in simulation and on neural recordings of hundreds of neurons across three brain areas. The frequency domain approach, in particular, consistently provided the greatest runtime benefits with the fewest trade-offs in statistical performance. We further characterize the estimation biases introduced by the frequency domain approach and demonstrate effective strategies to mitigate them. This work enables a powerful class of analysis techniques to keep pace with the growing scale of multi-population recordings, opening new avenues for exploring brain function.

Self-supervised vision-language models trained with contrastive objectives form the basis of current state-of-the-art methods in AI vision tasks. The success of these models is a direct consequence of the huge web-scale datasets used to train them, but they require correspondingly large vision components to properly learn powerful and general representations from such a broad data domain. This poses a challenge for deploying large vision-language models, especially in resource-constrained environments. To address this, we propose an alternate vision-language architecture, called HyperCLIP, that uses a small image encoder along with a hypernetwork that dynamically adapts image encoder weights to each new set of text inputs. All three components of the model (hypernetwork, image encoder, and text encoder) are pre-trained jointly end-to-end, and with a trained HyperCLIP model, we can generate new zero-shot deployment-friendly image classifiers for any task with a single forward pass through the text encoder and hypernetwork. HyperCLIP increases the zero-shot accuracy of SigLIP trained models with small image encoders by up to 3% on ImageNet and 5% on CIFAR-100 with minimal training throughput overhead.

Large Language Model (LLM) agents are increasingly being deployed as conversational assistants capable of performing complex real-world tasks through tool integration. This enhanced ability to interact with external systems and process various data sources, while powerful, introduces significant security vulnerabilities. In particular, indirect prompt injection attacks pose a critical threat, where malicious instructions embedded within external data sources can manipulate agents to deviate from user intentions. While existing defenses based on rule constraints, source spotlighting, and authentication protocols show promise, they struggle to maintain robust security while preserving task functionality. We propose a novel and orthogonal perspective that reframes agent security from preventing harmful actions to ensuring task alignment, requiring every agent action to serve user objectives. Based on this insight, we develop Task Shield, a test-time defense mechanism that systematically verifies whether each instruction and tool call contributes to user-specified goals. Through experiments on the AgentDojo benchmark, we demonstrate that Task Shield reduces attack success rates (2.07\%) while maintaining high task utility (69.79\%) on GPT-4o.

Reproducibility, endurance, driftless data retention, and fine resolution of the programmable conductance weights are key technological requirements against memristive artificial synapses in neural network applications. However, the inherent fluctuations in the active volume impose severe constraints on the weight resolution. In order to understand and push these limits, a comprehensive noise benchmarking and noise reduction protocol is introduced. Our approach goes beyond the measurement of steady-state readout noise levels and tracks the voltage-dependent noise characteristics all along the resistive switching $I(V)$ curves. Furthermore, we investigate the tunability of the noise level by dedicated voltage cycling schemes in our filamentary Ta$_2$O$_5$ memristors. This analysis highlights a broad, order-of-magnitude variability of the possible noise levels behind seemingly reproducible switching cycles. Our nonlinear noise spectroscopy measurements identify a subthreshold voltage region with voltage-boosted fluctuations. This voltage range enables the reconfiguration of the fluctuators without resistive switching, yielding a highly denoised state within a few subthreshold cycles., Comment: Main text 8 pages, 4 figures; ESI 9 pages, 5 figures

A generalization of Wilson line operators at subleading power in the soft expansion has been recently introduced as an efficient building block of gravitational scattering amplitudes for non-spinning objects. The classical limit in this picture corresponds to the strict Regge limit, where the Post-Minkowskian (PM) expansion corresponds to the soft expansion, interpreted as a sum over correlations of soft emissions. Building on the well-studied worldline model with ${\cal N}=1$ supersymmetry, in this work we extend the generalized Wilson line (GWL) approach to the case of spinning gravitating bodies. Specifically, at the quantum level we derive from first-principles a representation for the spin $1/2$ GWL that is relevant for the all-order factorization of next-to-soft gravitons with fermionic matter, thus generalizing the exponentiation of single-emission next-to-soft theorems. At the classical level, we identity the suitable generalization of Wilson line operators that enables the generation of classical spin observables at linear order in spin. Thanks to the crucial role played by the soft expansion, the map from Grassmann variables to classical spin is manifest. We also comment on the relation between the GWL approach and the Worldline Quantum Field Theory as well as the Heavy Mass Effective Theory formalism. We validate the approach by rederiving known results in the conservative sector at 2PM order., Comment: 48 pages

Network-based representations of fitness landscapes have grown in popularity in the past decade; this is probably because of growing interest in explainability for optimisation algorithms. Local optima networks (LONs) have been especially dominant in the literature and capture an approximation of local optima and their connectivity in the landscape. However, thus far, LONs have been constructed according to a strict definition of what a local optimum is: the result of local search. Many evolutionary approaches do not include this, however. Popular algorithms such as CMA-ES have therefore never been subject to LON analysis. Search trajectory networks (STNs) offer a possible alternative: nodes can be any search space location. However, STNs are not typically modelled in such a way that models temporal stalls: that is, a region in the search space where an algorithm fails to find a better solution over a defined period of time. In this work, we approach this by systematically analysing a special case of STN which we name attractor networks. These offer a coarse-grained view of algorithm behaviour with a singular focus on stall locations. We construct attractor networks for CMA-ES, differential evolution, and random search for 24 noiseless black-box optimisation benchmark problems. The properties of attractor networks are systematically explored. They are also visualised and compared to traditional LONs and STN models. We find that attractor networks facilitate insights into algorithm behaviour which other models cannot, and we advocate for the consideration of attractor analysis even for algorithms which do not include local search.

Image registration is fundamental in medical imaging applications, such as disease progression analysis or radiation therapy planning. The primary objective of image registration is to precisely capture the deformation between two or more images, typically achieved by minimizing an optimization problem. Due to its inherent ill-posedness, regularization is a key component in driving the solution toward anatomically meaningful deformations. A wide range of regularization methods has been proposed for both conventional and deep learning-based registration. However, the appropriate application of regularization techniques often depends on the specific registration problem, and no one-fits-all method exists. Despite its importance, regularization is often overlooked or addressed with default approaches, assuming existing methods are sufficient. A comprehensive and structured review remains missing. This review addresses this gap by introducing a novel taxonomy that systematically categorizes the diverse range of proposed regularization methods. It highlights the emerging field of learned regularization, which leverages data-driven techniques to automatically derive deformation properties from the data. Moreover, this review examines the transfer of regularization methods from conventional to learning-based registration, identifies open challenges, and outlines future research directions. By emphasizing the critical role of regularization in image registration, we hope to inspire the research community to reconsider regularization strategies in modern registration algorithms and to explore this rapidly evolving field further., Comment: Submitted to Medical Image Analysis

Despite the growing use of large language models (LLMs) for providing feedback, limited research has explored how to achieve high-quality feedback. This case study introduces an evaluation framework to assess different zero-shot prompt engineering methods. We varied the prompts systematically and analyzed the provided feedback on programming errors in R. The results suggest that prompts suggesting a stepwise procedure increase the precision, while omitting explicit specifications about which provided data to analyze improves error identification., Comment: 8 pages, 1 figure

Objective: The objective of this study is to develop and evaluate a systematic approach to optimize Deep Brain Stimulation (DBS) parameters, addressing the challenge of identifying patient-specific settings and optimal stimulation targets for various neurological and mental disorders. Methods: TuneS, a novel pipeline to predict clinically optimal DBS contact configurations based on predefined targets and constraints, is introduced. The method relies upon patient-specific models of stimulation spread and extends optimization beyond traditional neural structures to include automated, model-based targeting of streamlines. Results: Initial findings demonstrate that STN motor streamlines consistently receive a significant portion of the allocated stimulation volume, suggesting that a consistent portion of the stimulation should ideally focus on the STN motor streamlines. At the example of a small cohort of Parkinson's disease patients, the value of model-based contact predictions for assessing stimulation targets while observing constraints is demonstrated. Conclusion: TuneS shows promise as a research tool, enabling systematic assessment of DBS target effectiveness and facilitating constraint-aware optimization of stimulation parameters. Significance: The presented pipeline offers a pathway to improve patient-specific DBS therapies and contributes to the broader understanding of effective DBS targeting strategies., Comment: 8 pages, 9 figures, submitted to IEEE Transactions on Biomedical Engineering

We present observations of ASASSN-22ci (AT2022dbl), a nearby tidal disruption event (TDE) discovered by the All-Sky Automated Survey for Supernovae (ASAS-SN) at a distance of d$_L \simeq 125$ Mpc. Roughly two years after the initial ASAS-SN discovery, a second flare was detected coincident with ASASSN-22ci. UV/optical photometry and optical spectroscopy indicate that both flares are likely powered by TDEs. The striking similarity in flare properties suggests that these flares result from subsequent disruptions of the same star. Each flare rises on a timescale of $\sim$30 days, has a temperature of $\approx$30,000 K, a peak bolometric luminosity of $L_{UV/Opt} = 10^{43.6 - 43.9} \textrm{ erg} \textrm{ s}^{-1}$, and exhibits a blue optical spectrum with broad H, He, and N lines. No X-ray emission is detected during either flare, but X-ray emission with an unabsorbed luminosity of $L_{X} = 3\times10^{41} \textrm{ erg} \textrm{ s}^{-1}$ and $kT = 0.042$ eV is observed between the flares. Pre-discovery survey observations rule out the existence of earlier flares within the past $\approx$6000 days, indicating that the discovery of ASASSN-22ci likely coincides with the first flare. If the observed flare separation of $720 \pm 4.7$ days is the orbital period, the next flare of ASASSN-22ci should occur near MJD 61075 (2026 February 04). Finally, we find that the existing sample of repeating TDE candidates is consistent with Hills capture of a star initially in a binary with a total mass between $\sim$$1 - 4$ M$_{\odot}$ and a separation of $\sim$$0.01 - 0.1$ AU., Comment: 30 pages, 15 figures, 4 tables. Comments welcome

We argue that the Declarative Self-improving Python (DSPy) optimizers are a way to align the large language model (LLM) prompts and their evaluations to the human annotations. We present a comparative analysis of five teleprompter algorithms, namely, Cooperative Prompt Optimization (COPRO), Multi-Stage Instruction Prompt Optimization (MIPRO), BootstrapFewShot, BootstrapFewShot with Optuna, and K-Nearest Neighbor Few Shot, within the DSPy framework with respect to their ability to align with human evaluations. As a concrete example, we focus on optimizing the prompt to align hallucination detection (using LLM as a judge) to human annotated ground truth labels for a publicly available benchmark dataset. Our experiments demonstrate that optimized prompts can outperform various benchmark methods to detect hallucination, and certain telemprompters outperform the others in at least these experiments., Comment: 7 pages, 10 tables, two-column format

In quantum cryptography, fundamental laws of quantum physics are exploited to enhance the security of cryptographic tasks. Quantum key distribution is by far the most studied protocol to date, enabling the establishment of a secret key between trusted parties. However, there exist many practical use-cases in communication networks, which also involve parties in distrustful settings. The most fundamental quantum cryptographic building block in such a distrustful setting is quantum coin flipping, which provides an advantage compared to its classical equivalent. So far, few experimental studies on quantum coin flipping have been reported, all of which used probabilistic quantum light sources facing fundamental limitations. Here, we experimentally implement a quantum strong coin flipping protocol using single-photon states and demonstrate an advantage compared to both classical realizations and implementations using faint laser pulses. We achieve this by employing a state-of-the-art deterministic single-photon source based on the Purcell-enhanced emission of a semiconductor quantum dot in combination with fast polarization-state encoding enabling a quantum bit error ratio below 3%, required for the successful execution of the protocol. The reduced multi-photon emission yields a smaller bias of the coin flipping protocol compared to an attenuated laser implementation, both in simulations and in the experiment. By demonstrating a single-photon quantum advantage in a cryptographic primitive beyond QKD, our work represents a major advance towards the implementation of complex cryptographic tasks in a future quantum internet., Comment: 21 pages, 6 figures, 2 tables, 30 references

This summary of the second Terrestrial Very-Long-Baseline Atom Interferometry (TVLBAI) Workshop provides a comprehensive overview of our meeting held in London in April 2024, building on the initial discussions during the inaugural workshop held at CERN in March 2023. Like the summary of the first workshop, this document records a critical milestone for the international atom interferometry community. It documents our concerted efforts to evaluate progress, address emerging challenges, and refine strategic directions for future large-scale atom interferometry projects. Our commitment to collaboration is manifested by the integration of diverse expertise and the coordination of international resources, all aimed at advancing the frontiers of atom interferometry physics and technology, as set out in a Memorandum of Understanding signed by over 50 institutions., Comment: Summary of the second Terrestrial Very-Long-Baseline Atom Interferometry Workshop held at Imperial College London: https://indico.cern.ch/event/1369392/

We study stationary solutions of McKean-Vlasov equation on a high-dimensional sphere and other compact Riemannian manifolds. We extend the equivalence of the energetic problem formulation to the manifold setting and characterize critical points of the corresponding free energy functional. On a sphere, we employ the properties of spherical convolution to study the bifurcation branches around the uniform state. We also give a sufficient condition for an existence of a discontinuous transition point in terms of the interaction kernel and compare it to the Euclidean setting. We illustrate our results on a range of system, including the particle system arising from the transformer models and the Onsager model of liquid crystals., Comment: 56 pages

This paper studies how the Elo rating system behaves when the underlying modelling assumptions are not met., Comment: 29 pages

Unsupervised object-centric learning from videos is a promising approach to extract structured representations from large, unlabeled collections of videos. To support downstream tasks like autonomous control, these representations must be both compositional and temporally consistent. Existing approaches based on recurrent processing often lack long-term stability across frames because their training objective does not enforce temporal consistency. In this work, we introduce a novel object-level temporal contrastive loss for video object-centric models that explicitly promotes temporal consistency. Our method significantly improves the temporal consistency of the learned object-centric representations, yielding more reliable video decompositions that facilitate challenging downstream tasks such as unsupervised object dynamics prediction. Furthermore, the inductive bias added by our loss strongly improves object discovery, leading to state-of-the-art results on both synthetic and real-world datasets, outperforming even weakly-supervised methods that leverage motion masks as additional cues.

A family of generalized Korteweg-de Vries-Burgers equations in one space dimension with a nonlinear source is considered. The purpose of this contribution is twofold. On one hand, the local well-posedness of the Cauchy problem on periodic Sobolev spaces and the regularity of the data-solution map are established. On the other hand, it is proved that periodic traveling waves which are spectrally unstable are also orbitally (nonlinearly) unstable under the flow of the evolution equation in periodic Sobolev spaces with same period as the fundamental period of the wave. This orbital instability criterion hinges on the well-posedness of the Cauchy problem, on the smoothness of the data-solution map, as well as on an abstract result which provides sufficient conditions for the instability of equilibria under iterations of a nonlinear map in Banach spaces. Applications of the former criterion include the orbital instability of a family of small-amplitude periodic waves with finite fundamental period for the Korteweg-de Vries-Burgers-Fisher equation, which emerge from a local Hopf bifurcation around a critical value of the velocity., Comment: 24 pages

Gallium oxide is a novel advanced material gaining increasing attention for its unique combination of functional properties. It forms in several phases or polymorphs - {\alpha}, {\beta}, {\gamma}, and {\kappa} - having variable properties because of different lattice symmetries. Optical properties are of particular importance as they determine specific device applications and can also be used for phase identification. However, a direct comparison of optical polymorph signatures, including such critical parameters as bandgaps, is challenging due to the scattered, limited, or even absent data for certain phases in the literature. To address this issue, in the present work we systematically cross-correlate optical emission and absorption features of {\alpha}, {\beta}, {\gamma}, and {\kappa} thin films, as well as differently oriented {\beta}-phase bulk crystals and {\gamma}/{\beta} double polymorph structures. We demonstrate that the optical bandgap and emission features scale consistently across these polymorphs upon minimization of the methodological uncertainties. As a result, this work provides a comparative library of near- and far-field optical signatures of the polymorphs for use by the multidisciplinary research community working with gallium oxide., Comment: 17 pages, 14 figures, 4 tables

Far from equilibrium, universal dynamics prevails in many different situations, from pattern coarsening to turbulence, while the understanding of its microscopic laws and classification has remained unsatisfactory. Here, universal scaling evolution reflected by coarsening in a multi-component Bose gas is traced back to a low-energy effective theory given by a double sine-Gordon (DSG) model for a single real scalar field. Our experimental observations of a rubidium spinor BEC support the applicability of this model. Evaluating the scaling evolution according to the low-energy effective model demonstrates the universal aspects of the scaling characteristics as compared with that of the spinor gas. The difference between diffusion-type and sub-diffusive scaling is shown to be connected to the occupation of minima within the DSG potential. Our results point to a path towards a microscopic description and classification of pattern coarsening in many-body systems., Comment: 16pp., 8 figs., minor corrections

Disagreement in human labeling is ubiquitous, and can be captured in human judgment distributions (HJDs). Recent research has shown that explanations provide valuable information for understanding human label variation (HLV) and large language models (LLMs) can approximate HJD from a few human-provided label-explanation pairs. However, collecting explanations for every label is still time-consuming. This paper examines whether LLMs can be used to replace humans in generating explanations for approximating HJD. Specifically, we use LLMs as annotators to generate model explanations for a few given human labels. We test ways to obtain and combine these label-explanations with the goal to approximate human judgment distribution. We further compare the resulting human with model-generated explanations, and test automatic and human explanation selection. Our experiments show that LLM explanations are promising for NLI: to estimate HJD, generated explanations yield comparable results to human's when provided with human labels. Importantly, our results generalize from datasets with human explanations to i) datasets where they are not available and ii) challenging out-of-distribution test sets., Comment: 25 pages, 21 figures

Classifying scanned documents is a challenging problem that involves image, layout, and text analysis for document understanding. Nevertheless, for certain benchmark datasets, notably RVL-CDIP, the state of the art is closing in to near-perfect performance when considering hundreds of thousands of training samples. With the advent of large language models (LLMs), which are excellent few-shot learners, the question arises to what extent the document classification problem can be addressed with only a few training samples, or even none at all. In this paper, we investigate this question in the context of zero-shot prompting and few-shot model fine-tuning, with the aim of reducing the need for human-annotated training samples as much as possible., Comment: ICPR 2024