Novel architectures have recently improved generative image synthesis leading to excellent visual quality in various tasks. Of particular note is the field of ``AI-Art'', which has seen unprecedented growth with the emergence of powerful multimodal models such as CLIP. By combining speech and image synthesis models, so-called ``prompt-engineering'' has become established, in which carefully selected and composed sentences are used to achieve a certain visual style in the synthesized image. In this note, we present an alternative approach based on retrieval-augmented diffusion models (RDMs). In RDMs, a set of nearest neighbors is retrieved from an external database during training for each training instance, and the diffusion model is conditioned on these informative samples. During inference (sampling), we replace the retrieval database with a more specialized database that contains, for example, only images of a particular visual style. This provides a novel way to prompt a general trained model after training and thereby specify a particular visual style. As shown by our experiments, this approach is superior to specifying the visual style within the text prompt. We open-source code and model weights at this https URL .

Generative AI (GenAI) is increasingly used in survey contexts to simulate human preferences. While many research endeavors evaluate the quality of synthetic GenAI data by comparing model-generated responses to gold-standard survey results, fundamental questions about the validity and reliability of using LLMs as substitutes for human respondents remain. Our study provides a technical analysis of how demographic attributes and prompt variations influence latent opinion mappings in large language models (LLMs) and evaluates their suitability for survey-based predictions. Using 14 different models, we find that LLM-generated data fails to replicate the variance observed in real-world human responses, particularly across demographic subgroups. In the political space, persona-to-party mappings exhibit limited differentiation, resulting in synthetic data that lacks the nuanced distribution of opinions found in survey data. Moreover, we show that prompt sensitivity can significantly alter outputs for some models, further undermining the stability and predictiveness of LLM-based simulations. As a key contribution, we adapt a probe-based methodology that reveals how LLMs encode political affiliations in their latent space, exposing the systematic distortions introduced by these models. Our findings highlight critical limitations in AI-generated survey data, urging caution in its use for public opinion research, social science experimentation, and computational behavioral modeling.

Most machine learning algorithms are configured by one or several hyperparameters that must be carefully chosen and often considerably impact performance. To avoid a time consuming and unreproducible manual trial-and-error process to find well-performing hyperparameter configurations, various automatic hyperparameter optimization (HPO) methods, e.g., based on resampling error estimation for supervised machine learning, can be employed. After introducing HPO from a general perspective, this paper reviews important HPO methods such as grid or random search, evolutionary algorithms, Bayesian optimization, Hyperband and racing. It gives practical recommendations regarding important choices to be made when conducting HPO, including the HPO algorithms themselves, performance evaluation, how to combine HPO with ML pipelines, runtime improvements, and parallelization. This work is accompanied by an appendix that contains information on specific software packages in R and Python, as well as information and recommended hyperparameter search spaces for specific learning algorithms. We also provide notebooks that demonstrate concepts from this work as supplementary files.

Quantum emitters coupled to one-dimensional waveguides constitute a paradigmatic quantum-optical platform for exploring collective phenomena in open quantum many-body systems. For appropriately spaced emitters, they realize the Dicke model, whose characteristic permutation symmetry allows for efficient exact solutions featuring superradiance. When the emitters are arbitrarily spaced, however, this symmetry is lost and general analytical solutions are no longer available. In this work, we introduce a novel numerical method to study the dynamics of such systems by extending the time-dependent neural quantum state (t-NQS) framework to open quantum systems. We benchmark our approach across a range of waveguide QED settings and compare its performance with tensor-network calculations. Our results demonstrate that the t-NQS approach is competitive with other numerical methods and highlight the potential of t-NQSs for studying open quantum many-body systems out of equilibrium.

With the exponential growth of the life science literature, biomedical text mining (BTM) has become an essential technology for accelerating the extraction of insights from publications. Identifying named entities (e.g., diseases, drugs, or genes) in texts and their linkage to reference knowledge bases are crucial steps in BTM pipelines to enable information aggregation from different documents. However, tools for these two steps are rarely applied in the same context in which they were developed. Instead, they are applied in the wild, i.e., on application-dependent text collections different from those used for the tools' training, varying, e.g., in focus, genre, style, and text type. This raises the question of whether the reported performance of BTM tools can be trusted for downstream applications. Here, we report on the results of a carefully designed cross-corpus benchmark for named entity extraction, where tools were applied systematically to corpora not used during their training. Based on a survey of 28 published systems, we selected five for an in-depth analysis on three publicly available corpora encompassing four different entity types. Comparison between tools results in a mixed picture and shows that, in a cross-corpus setting, the performance is significantly lower than the one reported in an in-corpus setting. HunFlair2 showed the best performance on average, being closely followed by PubTator. Our results indicate that users of BTM tools should expect diminishing performances when applying them in the wild compared to original publications and show that further research is necessary to make BTM tools more robust.

Many applications require an understanding of an image that goes beyond the simple detection and classification of its objects. In particular, a great deal of semantic information is carried in the relationships between objects. We have previously shown that the combination of a visual model and a statistical semantic prior model can improve on the task of mapping images to their associated scene description. In this paper, we review the model and compare it to a novel conditional multi-way model for visual relationship detection, which does not include an explicitly trained visual prior model. We also discuss potential relationships between the proposed methods and memory models of the human brain.

Scientists and practitioners increasingly rely on machine learning to model data and draw conclusions. Compared to statistical modeling approaches, machine learning makes fewer explicit assumptions about data structures, such as linearity. However, their model parameters usually cannot be easily related to the data generating process. To learn about the modeled relationships, partial dependence (PD) plots and permutation feature importance (PFI) are often used as interpretation methods. However, PD and PFI lack a theory that relates them to the data generating process. We formalize PD and PFI as statistical estimators of ground truth estimands rooted in the data generating process. We show that PD and PFI estimates deviate from this ground truth due to statistical biases, model variance and Monte Carlo approximation errors. To account for model variance in PD and PFI estimation, we propose the learner-PD and the learner-PFI based on model refits, and propose corrected variance and confidence interval estimators.

The aim of this paper is to offer the first systematic exploration and definition of equivalent causal models in the context where both models are not made up of the same variables. The idea is that two models are equivalent when they agree on all "essential" causal information that can be expressed using their common variables. I do so by focussing on the two main features of causal models, namely their structural relations and their functional relations. In particular, I define several relations of causal ancestry and several relations of causal sufficiency, and require that the most general of these relations are preserved across equivalent models.

The aim of this paper is to offer the first systematic exploration and definition of equivalent causal models in the context where both models are not made up of the same variables. The idea is that two models are equivalent when they agree on all "essential" causal information that can be expressed using their common variables. I do so by focussing on the two main features of causal models, namely their structural relations and their functional relations. In particular, I define several relations of causal ancestry and several relations of causal sufficiency, and require that the most general of these relations are preserved across equivalent models.

The migratory dynamics of cells in physiological processes, ranging from wound healing to cancer metastasis, rely on contact-mediated cell-cell interactions. These interactions play a key role in shaping the stochastic trajectories of migrating cells. While data-driven physical formalisms for the stochastic migration dynamics of single cells have been developed, such a framework for the behavioral dynamics of interacting cells still remains elusive. Here, we monitor stochastic cell trajectories in a minimal experimental cell collider: a dumbbell-shaped micropattern on which pairs of cells perform repeated cellular collisions. We observe different characteristic behaviors, including cells reversing, following and sliding past each other upon collision. Capitalizing on this large experimental data set of coupled cell trajectories, we infer an interacting stochastic equation of motion that accurately predicts the observed interaction behaviors. Our approach reveals that interacting non-cancerous MCF10A cells can be described by repulsion and friction interactions. In contrast, cancerous MDA-MB-231 cells exhibit attraction and anti-friction interactions, promoting the predominant relative sliding behavior observed for these cells. Based on these experimentally inferred interactions, we show how this framework may generalize to provide a unifying theoretical description of the diverse cellular interaction behaviors of distinct cell types.

In this paper, we investigate the computational complexity of solutions to the Laplace and the diffusion equation. We show that for a certain class of initial-boundary value problems of the Laplace and the diffusion equation, the solution operator is $\# P_1/ \#P$-complete in the sense that it maps polynomial-time computable functions to the set of $\#P_1/ \#P$-complete functions. Consequently, there exists polynomial-time (Turing) computable input data such that the solution is not polynomial-time computable, unless $FP=\#P$ or $FP_1=\#P_1$. In this case, we can, in general, not simulate the solution of the Laplace or the diffusion equation on a digital computer without having a complexity blowup, i.e., the computation time for obtaining an approximation of the solution with up to a finite number of significant digits grows non-polynomially in the number of digits. This indicates that the computational complexity of the solution operator that models a physical phenomena is intrinsically high, independent of the numerical algorithm that is used to approximate a solution.

The pseudoinverse of a matrix, a generalized notion of the inverse, is of fundamental importance in linear algebra. However, there does not exist a closed form representation of the pseudoinverse, which can be straightforwardly computed. Therefore, an algorithmic computation is necessary. An algorithmic computation can only be evaluated by also considering the underlying hardware, typically digital hardware, which is responsible for performing the actual computations step by step. In this paper, we analyze if and to what degree the pseudoinverse actually can be computed on digital hardware platforms modeled as Turing machines. For this, we utilize the notion of an effective algorithm which describes a provably correct computation: upon an input of any error parameter, the algorithm provides an approximation within the given error bound with respect to the unknown solution. We prove that an effective algorithm for computing the pseudoinverse of any matrix can not exist on a Turing machine, although provably correct algorithms do exist for specific classes of matrices. Even more, our results introduce a lower bound on the accuracy that can be obtained algorithmically when computing the pseudoinverse on Turing machines.

Quantum computing promises potential for science and industry by solving certain computationally complex problems faster than classical computers. Quantum computing systems evolved from monolithic systems towards modular architectures comprising multiple quantum processing units (QPUs) coupled to classical computing nodes (HPC). With the increasing scale, middleware systems that facilitate the efficient coupling of quantum-classical computing are becoming critical. Through an in-depth analysis of quantum applications, integration patterns and systems, we identified a gap in understanding Quantum-HPC middleware systems. We present a conceptual middleware to facilitate reasoning about quantum-classical integration and serve as the basis for a future middleware system. An essential contribution of this paper lies in leveraging well-established high-performance computing abstractions for managing workloads, tasks, and resources to integrate quantum computing into HPC systems seamlessly.

The interplay between bacterial chromosome organization and functions such as transcription and replication can be studied in increasing detail using novel experimental techniques. Interpreting the resulting quantitative data, however, can be theoretically challenging. In this minireview, we discuss how connecting experimental observations to biophysical theory and modeling can give rise to new insights on bacterial chromosome organization. We consider three flavors of models of increasing complexity: simple polymer models that explore how physical constraints, such as confinement or plectoneme branching, can affect bacterial chromosome organization; bottom-up mechanistic models that connect these constraints to their underlying causes, for instance chromosome compaction to macromolecular crowding, or supercoiling to transcription; and finally, data-driven methods for inferring interpretable and quantitative models directly from complex experimental data. Using recent examples, we discuss how biophysical models can both deepen our understanding of how bacterial chromosomes are structured, and give rise to novel predictions about bacterial chromosome organization.

Non-equilibrium selection pressures were proposed for the formation of oligonucleotides with rich functionalities encoded in their sequences, such as catalysis. Since phase separation was shown to direct various chemical processes, we ask whether condensed phases can provide mechanisms for sequence selection. To answer this question, we use non-equilibrium thermodynamics and describe the reversible oligomerization of different monomers to sequences at non-dilute conditions prone to phase separation. We find that when sequences oligomerize, their interactions give rise to phase separation, boosting specific sequences' enrichment and depletion. Our key result is that phase separation gives rise to a selection pressure for the oligomerization of specific sequence patterns when fragmentation maintains the system away from equilibrium. Specifically, slow fragmentation favors alternating sequences that interact well with their environment (more cooperative), while fast fragmentation selects sequences with extended motifs capable of specific sequence interactions (less cooperative). Our results highlight that out-of-equilibrium condensed phases could provide versatile hubs for Darwinian-like evolution toward functional sequences, both relevant for the molecular origin of life and de novo life.

Schrödinger connections are a special class of affine connections, which despite being metric incompatible, preserve length of vectors under autoparallel transport. In the present paper, we introduce a novel coordinate-free formulation of Schrödinger connections. After recasting their basic properties in the language of differential geometry, we show that Schrödinger connections can be realized through torsion, non-metricity, or both. We then calculate the curvature tensors of Yano-Schrödinger geometry and present the first explicit example of a non-static Einstein manifold with torsion. We generalize the Raychaudhuri and Sachs equations to the Schrödinger geometry. The length-preserving property of these connections enables us to construct a Lagrangian formulation of the Sachs equation. We also obtain an equation for cosmological distances. After this geometric analysis, we build gravitational theories based on Yano-Schrödinger geometry, using both a metric and a metric-affine approach. For the latter, we introduce a novel cosmological hyperfluid that will source the Schrödinger geometry. Finally, we construct simple cosmological models within these theories and compare our results with observational data as well as the $\Lambda$CDM model.

Representing the glaciation of mixed-phase clouds in terms of the Wegener-Bergeron-Findeisen process is a challenge for many weather and climate models, which tend to overestimate this process because cloud dynamics and microphysics are not accurately represented. As turbulence is essential for the transport of water vapour from evaporating liquid droplets to ice crystals, we developed a statistical model using established closures to assess the role of small-scale turbulence. The model successfully captures results of direct numerical simulations, and we use it to assess the role of small-scale turbulence. We find that small-scale turbulence broadens the droplet-size distribution somewhat, but it does not significantly affect the glaciation time on submetre scales. However, our analysis indicates that turbulence on larger spatial scales is likely to affect ice growth. While the model must be amended to describe larger scales, the present work facilitates a path forward to understanding the role of turbulence in the Wegener-Bergeron-Findeisen process.

Despite the ubiquity of large language models (LLMs) in AI research, the question of embodiment in LLMs remains underexplored, distinguishing them from embodied systems in robotics where sensory perception directly informs physical action. Our investigation navigates the intriguing terrain of whether LLMs, despite their non-embodied nature, effectively capture implicit human intuitions about fundamental, spatial building blocks of language. We employ insights from spatial cognitive foundations developed through early sensorimotor experiences, guiding our exploration through the reproduction of three psycholinguistic experiments. Surprisingly, correlations between model outputs and human responses emerge, revealing adaptability without a tangible connection to embodied experiences. Notable distinctions include polarized language model responses and reduced correlations in vision language models. This research contributes to a nuanced understanding of the interplay between language, spatial experiences, and the computations made by large language models. More at this https URL

Coherent Diffraction Imaging (CDI) is an experimental technique to gain images of isolated structures by recording the light scattered off the sample. In principle, the sample density can be recovered from the scattered light field through a straightforward Fourier Transform operation. However, only the amplitude of the field is recorded, while the phase is lost during the measurement process and has to be retrieved by means of suitable, well-established phase retrieval algorithms. In this work, we present SPRING, an analysis framework tailored to X-ray Free Electron Laser (XFEL) single-shot single-particle diffraction data that implements the Memetic Phase Retrieval method to mitigate the shortcomings of conventional algorithms. We benchmark the approach on experimental data acquired in two experimental campaigns at SwissFEL and European XFEL. Imaging results on isolated nanostructures reveal unprecedented stability and resilience of the algorithm's behavior on the input parameters, as well as the capability of identifying the solution in conditions hardly treatable so far with conventional methods. A user-friendly implementation of SPRING is released as open-source software, aiming at being a reference tool for the coherent diffraction imaging community at XFEL and synchrotron facilities.