Search

The field of Explainable Artificial Intelligence (XAI) for Deep Neural Network models has developed significantly, offering numerous techniques to extract explanations from models. However, evaluating explanations is often not trivial, and differences in applied metrics can be subtle, especially with non-intelligible data. Thus, there is a need for visualizations tailored to explore explanations for domains with such data, e.g., time series. We propose DAVOTS, an interactive visual analytics approach to explore raw time series data, activations of neural networks, and attributions in a dense-pixel visualization to gain insights into the data, models' decisions, and explanations. To further support users in exploring large datasets, we apply clustering approaches to the visualized data domains to highlight groups and present ordering strategies for individual and combined data exploration to facilitate finding patterns. We visualize a CNN trained on the FordA dataset to demonstrate the approach., Comment: 5 pages, 2 figures, accepted at MLVIS 2023

We propose an interactive methodology for generating counterfactual explanations for univariate time series data in classification tasks by leveraging 2D projections and decision boundary maps to tackle interpretability challenges. Our approach aims to enhance the transparency and understanding of deep learning models' decision processes. The application simplifies the time series data analysis by enabling users to interactively manipulate projected data points, providing intuitive insights through inverse projection techniques. By abstracting user interactions with the projected data points rather than the raw time series data, our method facilitates an intuitive generation of counterfactual explanations. This approach allows for a more straightforward exploration of univariate time series data, enabling users to manipulate data points to comprehend potential outcomes of hypothetical scenarios. We validate this method using the ECG5000 benchmark dataset, demonstrating significant improvements in interpretability and user understanding of time series classification. The results indicate a promising direction for enhancing explainable AI, with potential applications in various domains requiring transparent and interpretable deep learning models. Future work will explore the scalability of this method to multivariate time series data and its integration with other interpretability techniques., Comment: 14 pages, 4 figures, accepted at XKDD @ ECML-PKDD

Understanding how models process and interpret time series data remains a significant challenge in deep learning to enable applicability in safety-critical areas such as healthcare. In this paper, we introduce Sequence Dreaming, a technique that adapts Activation Maximization to analyze sequential information, aiming to enhance the interpretability of neural networks operating on univariate time series. By leveraging this method, we visualize the temporal dynamics and patterns most influential in model decision-making processes. To counteract the generation of unrealistic or excessively noisy sequences, we enhance Sequence Dreaming with a range of regularization techniques, including exponential smoothing. This approach ensures the production of sequences that more accurately reflect the critical features identified by the neural network. Our approach is tested on a time series classification dataset encompassing applications in predictive maintenance. The results show that our proposed Sequence Dreaming approach demonstrates targeted activation maximization for different use cases so that either centered class or border activation maximization can be generated. The results underscore the versatility of Sequence Dreaming in uncovering salient temporal features learned by neural networks, thereby advancing model transparency and trustworthiness in decision-critical domains., Comment: 16 pages, 4 figures, accepted at TempXAI @ ECML-PKDD

Efficient public transport systems are crucial for sustainable urban development as cities face increasing mobility demands. Yet, many public transport networks struggle to meet diverse user needs due to historical development, urban constraints, and financial limitations. Traditionally, planning of transport network structure is often based on limited surveys, expert opinions, or partial usage statistics. This provides an incomplete basis for decision-making. We introduce an data-driven approach to public transport planning and optimization, calculating detailed accessibility measures at the individual housing level. Our visual analytics workflow combines population-group-based simulations with dynamic infrastructure analysis, utilizing a scenario-based model to simulate daily travel patterns of varied demographic groups, including schoolchildren, students, workers, and pensioners. These population groups, each with unique mobility requirements and routines, interact with the transport system under different scenarios traveling to and from Points of Interest (POI), assessed through travel time calculations. Results are visualized through heatmaps, density maps, and network overlays, as well as detailed statistics. Our system allows us to analyze both the underlying data and simulation results on multiple levels of granularity, delivering both broad insights and granular details. Case studies with the city of Konstanz, Germany reveal key areas where public transport does not meet specific needs, confirmed through a formative user study. Due to the high cost of changing legacy networks, our analysis facilitates the identification of strategic enhancements, such as optimized schedules or rerouting, and few targeted stop relocations, highlighting consequential variations in accessibility to pinpointing critical service gaps., Comment: 9 pages, 8 figures

In academic research, systematic literature reviews are foundational and highly relevant, yet tedious to create due to the high volume of publications and labor-intensive processes involved. Systematic selection of relevant papers through conventional means like keyword-based filtering techniques can sometimes be inadequate, plagued by semantic ambiguities and inconsistent terminology, which can lead to sub-optimal outcomes. To mitigate the required extensive manual filtering, we explore and evaluate the potential of using Large Language Models (LLMs) to enhance the efficiency, speed, and precision of literature review filtering, reducing the amount of manual screening required. By using models as classification agents acting on a structured database only, we prevent common problems inherent in LLMs, such as hallucinations. We evaluate the real-world performance of such a setup during the construction of a recent literature survey paper with initially more than 8.3k potentially relevant articles under consideration and compare this with human performance on the same dataset. Our findings indicate that employing advanced LLMs like GPT-4o, Claude 3.5 Sonnet, Gemini 1.5 Flash, or Llama3 with simple prompting can significantly reduce the time required for literature filtering - from usually weeks of manual research to only a few minutes. Simultaneously, we crucially show that false negatives can indeed be controlled through a consensus scheme, achieving recalls >98.8% at or even beyond the typical human error threshold, thereby also providing for more accurate and relevant articles selected. Our research not only demonstrates a substantial improvement in the methodology of literature reviews but also sets the stage for further integration and extensive future applications of responsible AI in academic research practices., Comment: 5 pages, 5 figures, 1 table

Manual melody detection is a tedious task requiring high expertise level, while automatic detection is often not expressive or powerful enough. Thus, we present MelodyVis, a visual application designed in collaboration with musicology experts to explore melodic patterns in digital sheet music. MelodyVis features five connected views, including a Melody Operator Graph and a Voicing Timeline. The system utilizes eight atomic operators, such as transposition and mirroring, to capture melody repetitions and variations. Users can start their analysis by manually selecting patterns in the sheet view, and then identifying other patterns based on the selected samples through an interactive exploration process. We conducted a user study to investigate the effectiveness and usefulness of our approach and its integrated melodic operators, including usability and mental load questions. We compared the analysis executed by 25 participants with and without the operators. The study results indicate that the participants could identify at least twice as many patterns with activated operators. MelodyVis allows analysts to steer the analysis process and interpret results. Our study also confirms the usefulness of MelodyVis in supporting common analytical tasks in melodic analysis, with participants reporting improved pattern identification and interpretation. Thus, MelodyVis addresses the limitations of fully-automated approaches, enabling music analysts to step into the analysis process and uncover and understand intricate melodic patterns and transformations in sheet music., Comment: 9+2 pages, 9 figures, preprint, originally submitted to IEEE VIS 23, revision

The Dragonfly Telephoto Array employs a unique design to detect very large and diffuse galaxies, which might be missed with conventional telescopes. The Dragonfly Ultrawide Survey (DFUWS) is a new wide-field survey which will cover 10,000 deg$^2$ of the northern sky, and it provides an ideal dataset to find these large diffuse galaxies. From 3100 deg$^2$ of DFUWS data, we identified eleven large, low surface brightness galaxies as a pilot sample for spectroscopic follow-up. These are the largest galaxies in the examined area that appear smooth and isolated, with effective radii of 12"-27". Eight are below 24 $\mathrm{mag\,arcsec^{-2}}$ in central $g$-band surface brightness. Keck Cosmic Web Imager (KCWI) spectra of the diffuse light show that all eleven galaxies in this sample are quiescent, and seven qualify as ultra-diffuse galaxies (UDGs). Eight galaxies have distances between 15 and 30 Mpc, while the other three are in the Pegasus cluster at 50 Mpc. Their spectra show evidence of a $\sim 1$Gyr old stellar population in addition to an even older stellar population. The intermediate-age component is present in group and satellite galaxies but not in the Pegasus cluster UDGs. All galaxies in this sample are detected in both Dragonfly and Legacy imaging, and the sample partially overlaps with existing UDG catalogs. This pilot sample provides an excellent training set for our analysis of the upcoming full 10,000 deg$^2$ DFUWS data, from which we may expect to discover even larger, previously-unknown galaxies., Comment: Submitted to ApJ

Solids are rigid, which means that when left undisturbed, their structures are nearly static. It follows that these structures depend on history -- but it is surprising that they hold readable memories of past events. Here we review the research that has recently flourished around mechanical memory formation, beginning with amorphous solids' various memories of deformation and mesoscopic models based on particle rearrangements. We describe how these concepts apply to a much wider range of solids and glassy matter -- and how they are a bridge to memory and physical computing in mechanical metamaterials. An understanding of memory in all these solids can potentially be the basis for designing or training functionality into materials. Just as important is memory's value for understanding matter whenever it is complex, frustrated, and out of equilibrium., Comment: 22 pages, 7 figures. Invited submission to Annual Review of Condensed Matter Physics, Vol 16

Topological defects are singularities in the order parameter space that are mathematically described by topological invariants and cannot be removed by continuous transformations. These defects play a significant role in various fields, ranging from cosmology to solid-state physics and biological matter. The definition of these irregularities requires an ordered reference configuration, leading to decades of debate about the existence of topological defects in disordered systems, such as glasses. Recently, it has been proposed that well-defined topological defects might emerge in the dynamical properties of glasses under deformation, potentially relating to their plastic behavior. In this study, we investigate a two-dimensional colloidal glass system composed of particles interacting by an effective magnetic repulsion. We reveal the presence of topological defects in the eigenspace of the vibrational frequencies of this experimental 2D amorphous solid. The vibrational density of states of this 2D amorphous material exhibits distinct glassy properties such as the presence of a boson peak anomaly. Through extensive numerical and theoretical quantitative analysis, we establish a robust positive correlation between the vibrational characteristics and the total number of topological defects. We furthermore show that defects of opposite charge tend to pair together and prove their local statistical correlation with the "soft spots", the regions more prone to plastic flow. This work provides the experimental confirmation for the existence of topological defects in disordered systems revealing a complex interplay between topology, disorder, and vibrational behavior.

This work presents GAL{\AE}XI as a novel, energy-efficient flow solver for the simulation of compressible flows on unstructured meshes leveraging the parallel computing power of modern Graphics Processing Units (GPUs). GAL{\AE}XI implements the high-order Discontinuous Galerkin Spectral Element Method (DGSEM) using shock capturing with a finite-volume subcell approach to ensure the stability of the high-order scheme near shocks. This work provides details on the general code design, the parallelization strategy, and the implementation approach for the compute kernels with a focus on the element local mappings between volume and surface data due to the unstructured mesh. GAL{\AE}XI exhibits excellent strong scaling properties up to 1024 GPUs if each GPU is assigned a minimum of one million degrees of freedom degrees of freedom. To verify its implementation, a convergence study is performed that recovers the theoretical order of convergence of the implemented numerical schemes. Moreover, the solver is validated using both the incompressible and compressible formulation of the Taylor-Green-Vortex at a Mach number of 0.1 and 1.25, respectively. A mesh convergence study shows that the results converge to the high-fidelity reference solution and that the results match the original CPU implementation. Finally, GAL{\AE}XI is applied to a large-scale wall-resolved large eddy simulation of a linear cascade of the NASA Rotor 37. Here, the supersonic region and shocks at the leading edge are captured accurately and robustly by the implemented shock-capturing approach. It is demonstrated that GAL{\AE}XI requires less than half of the energy to carry out this simulation in comparison to the reference CPU implementation. This renders GAL{\AE}XI as a potent tool for accurate and efficient simulations of compressible flows in the realm of exascale computing and the associated new HPC architectures., Comment: 19 pages, 12 figures, 3 tables. Code available at: https://github.com/flexi-framework/galaexi

Categorical data does not have an intrinsic definition of distance or order, and therefore, established visualization techniques for categorical data only allow for a set-based or frequency-based analysis, e.g., through Euler diagrams or Parallel Sets, and do not support a similarity-based analysis. We present a novel dimensionality reduction-based visualization for categorical data, which is based on defining the distance of two data items as the number of varying attributes. Our technique enables users to pre-attentively detect groups of similar data items and observe the properties of the projection, such as attributes strongly influencing the embedding. Our prototype visually encodes data properties in an enhanced scatterplot-like visualization, encoding attributes in the background to show the distribution of categories. In addition, we propose two graph-based measures to quantify the plot's visual quality, which rank attributes according to their contribution to cluster cohesion. To demonstrate the capabilities of our similarity-based approach, we compare it to Euler diagrams and Parallel Sets regarding visual scalability and show its benefits through an expert study with five data scientists analyzing the Titanic and Mushroom datasets with up to 23 attributes and 8124 category combinations. Our results indicate that the Categorical Data Map offers an effective analysis method, especially for large datasets with a high number of category combinations., Comment: Fully replaced; 10 pages, 9 figures, LaTeX; to appear at Visual Data Science (VDS) Symposium at IEEE VIS 2024

In photonic crystals the propagation of light is governed by their photonic band structure, an ensemble of propagating states grouped into bands, separated by photonic band gaps. Due to discrete symmetries in spatially strictly periodic dielectric structures their photonic band structure is intrinsically anisotropic. However, for many applications, such as manufacturing artificial structural color materials or developing photonic computing devices, but also for the fundamental understanding of light-matter interactions, it is of major interest to seek materials with long range non-periodic dielectric structures which allow the formation of {\it isotropic} photonic band gaps. Here, we report the first ever 3D isotropic photonic band gap for an optimized disordered stealthy hyperuniform structure for microwaves. The transmission spectra are directly compared to a diamond pattern and an amorphous structure with similar node density. The band structure is measured experimentally for all three microwave structures, manufactured by 3D-Laser-printing for meta-materials with refractive index up to $n=2.1$. Results agree well with finite-difference-time-domain numerical investigations and a priori calculations of the band-gap for the hyperuniform structure: the diamond structure shows gaps but being anisotropic as expected, the stealthy hyperuniform pattern shows an isotropic gap of very similar magnitude, while the amorphous structure does not show a gap at all. The centimeter scaled microwave structures may serve as prototypes for micrometer scaled structures with bandgaps in the technologically very interesting region of infrared (IR)., Comment: 9 pages, 6 figures

Large language models (LLMs) are widely deployed in various downstream tasks, e.g., auto-completion, aided writing, or chat-based text generation. However, the considered output candidates of the underlying search algorithm are under-explored and under-explained. We tackle this shortcoming by proposing a tree-in-the-loop approach, where a visual representation of the beam search tree is the central component for analyzing, explaining, and adapting the generated outputs. To support these tasks, we present generAItor, a visual analytics technique, augmenting the central beam search tree with various task-specific widgets, providing targeted visualizations and interaction possibilities. Our approach allows interactions on multiple levels and offers an iterative pipeline that encompasses generating, exploring, and comparing output candidates, as well as fine-tuning the model based on adapted data. Our case study shows that our tool generates new insights in gender bias analysis beyond state-of-the-art template-based methods. Additionally, we demonstrate the applicability of our approach in a qualitative user study. Finally, we quantitatively evaluate the adaptability of the model to few samples, as occurring in text-generation use cases., Comment: 24 pages paper, 4 pages references, 3 pages appendix, 8 figures

Scatter plots are popular for displaying 2D data, but in practice, many data sets have more than two dimensions. For the analysis of such multivariate data, it is often necessary to switch between scatter plots of different dimension pairs, e.g., in a scatter plot matrix (SPLOM). Alternative approaches include a "grand tour" for an overview of the entire data set or creating artificial axes from dimensionality reduction (DR). A cross-cutting concern in all techniques is the ability of viewers to find correspondence between data points in different views. Previous work proposed animations to preserve the mental map between view changes and to trace points as well as clusters between scatter plots of the same underlying data set. In this paper, we evaluate a variety of spline- and rotation-based view transitions in a crowdsourced user study focusing on ecological validity. Using the study results, we assess each animation's suitability for tracing points and clusters across view changes. We evaluate whether the order of horizontal and vertical rotation is relevant for task accuracy. The results show that rotations with an orthographic camera or staged expansion of a depth axis significantly outperform all other animation techniques for the traceability of individual points. Further, we provide a ranking of the animated transition techniques for traceability of individual points. However, we could not find any significant differences for the traceability of clusters. Furthermore, we identified differences by animation direction that could guide further studies to determine potential confounds for these differences. We publish the study data for reuse and provide the animation framework as a D3.js plug-in.

AI-driven models are increasingly deployed in operational analytics solutions, for instance, in investigative journalism or the intelligence community. Current approaches face two primary challenges: ethical and privacy concerns, as well as difficulties in efficiently combining heterogeneous data sources for multimodal analytics. To tackle the challenge of multimodal analytics, we present MULTI-CASE, a holistic visual analytics framework tailored towards ethics-aware and multimodal intelligence exploration, designed in collaboration with domain experts. It leverages an equal joint agency between human and AI to explore and assess heterogeneous information spaces, checking and balancing automation through Visual Analytics. MULTI-CASE operates on a fully-integrated data model and features type-specific analysis with multiple linked components, including a combined search, annotated text view, and graph-based analysis. Parts of the underlying entity detection are based on a RoBERTa-based language model, which we tailored towards user requirements through fine-tuning. An overarching knowledge exploration graph combines all information streams, provides in-situ explanations, transparent source attribution, and facilitates effective exploration. To assess our approach, we conducted a comprehensive set of evaluations: We benchmarked the underlying language model on relevant NER tasks, achieving state-of-the-art performance. The demonstrator was assessed according to intelligence capability assessments, while the methodology was evaluated according to ethics design guidelines. As a case study, we present our framework in an investigative journalism setting, supporting war crime investigations. Finally, we conduct a formative user evaluation with domain experts in law enforcement. Our evaluations confirm that our framework facilitates human agency and steering in security-sensitive applications., Comment: 6 pages, 3 figures, 1 table

The growing popularity of generative language models has amplified interest in interactive methods to guide model outputs. Prompt refinement is considered one of the most effective means to influence output among these methods. We identify several challenges associated with prompting large language models, categorized into data- and model-specific, linguistic, and socio-linguistic challenges. A comprehensive examination of model outputs, including runner-up candidates and their corresponding probabilities, is needed to address these issues. The beam search tree, the prevalent algorithm to sample model outputs, can inherently supply this information. Consequently, we introduce an interactive visual method for investigating the beam search tree, facilitating analysis of the decisions made by the model during generation. We quantitatively show the value of exposing the beam search tree and present five detailed analysis scenarios addressing the identified challenges. Our methodology validates existing results and offers additional insights., Comment: 9 pages paper, 2 pages references, 7 figures

Given the increasing amount and general complexity of time series data in domains such as finance, weather forecasting, and healthcare, there is a growing need for state-of-the-art performance models that can provide interpretable insights into underlying patterns and relationships. Attribution techniques enable the extraction of explanations from time series models to gain insights but are hard to evaluate for their robustness and trustworthiness. We propose the Attribution Stability Indicator (ASI), a measure to incorporate robustness and trustworthiness as properties of attribution techniques for time series into account. We extend a perturbation analysis with correlations of the original time series to the perturbed instance and the attributions to include wanted properties in the measure. We demonstrate the wanted properties based on an analysis of the attributions in a dimension-reduced space and the ASI scores distribution over three whole time series classification datasets., Comment: 16 pages, 6 figures, ECML-PKDD Workshop XAI-TS: Explainable AI for Time Series: Advances and Applications

The Laser Interferometer Space Antenna (LISA) is expected to detect a wide variety of gravitational wave sources in the mHz band. Some of these signals will elude individual detection, instead contributing as confusion noise to one of several stochastic gravitational-wave backgrounds (SGWBs) -- notably including the `Galactic foreground', a loud signal resulting from the superposition of millions of unresolved double white dwarf binaries (DWDs) in the Milky Way. It is possible that similar, weaker SGWBs will be detectable from other DWD populations in the local universe, including the Large Magellanic Cloud (LMC). We use the Bayesian LISA Inference Package ($\tt{BLIP}$) to investigate the possibility of an anisotropic SGWB generated by unresolved DWDs in the LMC. To do so, we compute the LMC SGWB from a realistic DWD population generated via binary population synthesis, simulate four years of time-domain data with $\tt{BLIP}$ comprised of stochastic contributions from the LMC SGWB and the LISA detector noise, and analyze this data with $\tt{BLIP}$'s spherical harmonic anisotropic SGWB search. We also consider the case of spectral separation from the Galactic foreground. We present the results of these analyses and show, for the first time, that the unresolved DWDs in the LMC will comprise a significant SGWB for LISA., Comment: 11 pages, 8 figures. Published in MNRAS May 2024

To use reinforcement learning from human feedback (RLHF) in practical applications, it is crucial to learn reward models from diverse sources of human feedback and to consider human factors involved in providing feedback of different types. However, the systematic study of learning from diverse types of feedback is held back by limited standardized tooling available to researchers. To bridge this gap, we propose RLHF-Blender, a configurable, interactive interface for learning from human feedback. RLHF-Blender provides a modular experimentation framework and implementation that enables researchers to systematically investigate the properties and qualities of human feedback for reward learning. The system facilitates the exploration of various feedback types, including demonstrations, rankings, comparisons, and natural language instructions, as well as studies considering the impact of human factors on their effectiveness. We discuss a set of concrete research opportunities enabled by RLHF-Blender. More information is available at https://rlhfblender.info/., Comment: 14 pages, 3 figures

Optimal ski route selection is a challenge based on a multitude of factors, such as the steepness, compass direction, or crowdedness. The personal preferences of every skier towards these factors require individual adaptations, which aggravate this task. Current approaches within this domain do not combine automated routing capabilities with user preferences, missing out on the possibility of integrating domain knowledge in the analysis process. We introduce SkiVis, a visual analytics application to interactively explore ski slopes and provide routing recommendations based on user preferences. In collaboration with ski guides and enthusiasts, we elicited requirements and guidelines for such an application and propose different workflows depending on the skiers' familiarity with the resort. In a case study on the resort of Ski Arlberg, we illustrate how to leverage volunteered geographic information to enable a numerical comparison between slopes. We evaluated our approach through a pair-analytics study and demonstrate how it supports skiers in discovering relevant and preference-based ski routes. Besides the tasks investigated in the study, we derive additional use cases from the interviews that showcase the further potential of SkiVis, and contribute directions for further research opportunities., Comment: 11 pages, 10 figures

With the rising necessity of explainable artificial intelligence (XAI), we see an increase in task-dependent XAI methods on varying abstraction levels. XAI techniques on a global level explain model behavior and on a local level explain sample predictions. We propose a visual analytics workflow to support seamless transitions between global and local explanations, focusing on attributions and counterfactuals on time series classification. In particular, we adapt local XAI techniques (attributions) that are developed for traditional datasets (images, text) to analyze time series classification, a data type that is typically less intelligible to humans. To generate a global overview, we apply local attribution methods to the data, creating explanations for the whole dataset. These explanations are projected onto two dimensions, depicting model behavior trends, strategies, and decision boundaries. To further inspect the model decision-making as well as potential data errors, a what-if analysis facilitates hypothesis generation and verification on both the global and local levels. We constantly collected and incorporated expert user feedback, as well as insights based on their domain knowledge, resulting in a tailored analysis workflow and system that tightly integrates time series transformations into explanations. Lastly, we present three use cases, verifying that our technique enables users to (1)~explore data transformations and feature relevance, (2)~identify model behavior and decision boundaries, as well as, (3)~the reason for misclassifications., Comment: 14 pages, 6 figures, submitted to IEEE TVCG in December

Explainable Artificial Intelligence (XAI) has gained significant attention recently as the demand for transparency and interpretability of machine learning models has increased. In particular, XAI for time series data has become increasingly important in finance, healthcare, and climate science. However, evaluating the quality of explanations, such as attributions provided by XAI techniques, remains challenging. This paper provides an in-depth analysis of using perturbations to evaluate attributions extracted from time series models. A perturbation analysis involves systematically modifying the input data and evaluating the impact on the attributions generated by the XAI method. We apply this approach to several state-of-the-art XAI techniques and evaluate their performance on three time series classification datasets. Our results demonstrate that the perturbation analysis approach can effectively evaluate the quality of attributions and provide insights into the strengths and limitations of XAI techniques. Such an approach can guide the selection of XAI methods for time series data, e.g., focusing on return time rather than precision, and facilitate the development of more reliable and interpretable machine learning models for time series analysis., Comment: 16 pages, 2 pages references, 5 figures, 3 tables, submitted and accepted at xAI 2023

Unbiased sky background modeling is crucial for the analysis of deep wide-field images, but it remains a major challenge in low surface brightness astronomy. Traditional image processing algorithms are often designed to produce artificially flat backgrounds, erasing astrophysically meaningful structures. In this paper, we present three ideas that can be combined to produce wide-field astronomical data that preserve accurate representations of the background sky: (1) Use of all-sky infrared/sub-mm data to remove the large-scale time-varying components while leaving the scattered light from Galactic cirrus intact, with the assumptions of (a) the underlying background has little power on small scales, and (b) the Galactic cirrus in the field is optically thin on large scales; (2) Censoring of frames contaminated by anomalously prominent wings in the wide-angle point-spread function; and (3) Incorporation of spatial covariance in image stacking that controls the local background consistency. We demonstrate these methods using example datasets obtained with the Dragonfly Telephoto Array, but these general techniques are prospective to be applied to improve sky models in data obtained from other wide-field imaging surveys, including those from the upcoming Vera Rubin Telescope., Comment: 28 pages, 17 figures, accepted for publication in ApJ

Studies of glassy systems have shown how cyclic driving forms memories of amplitude. We explore how choice of driving protocol reveals dramatically different features of this memory. We model rearranging soft spots in sheared amorphous solids as hysterons. Cyclic shear with positive and negative shear strain reveals a return-point memory of multiple strains known from experiments and molecular dynamics simulations, while asymmetric driving (e.g. only positive shear strains) suppresses multiple memories. However, when we introduce frustrated interactions between hysterons, we identify a different mechanism that restores multiple memories for asymmetric driving and can be used for design. Our work suggests that this enhanced memory is a signature of frustration., Comment: 5 pages, 5 figures

A buckled sheet offers a reservoir of material that can be unfurled at a later time. For sufficiently thin yet stiff materials, this geometric process has a striking mechanical feature: when the slack runs out, the material locks to further extension. Here we establish a simple route to a tunable locking material - a system with an interval where it is freely deformable under a given deformation mode, and where the endpoints of this interval can be changed continuously over a wide range. We demonstrate this type of mechanical response in a thin cylindrical shell subjected to axial twist and compression, and we rationalize our results with a simple geometric model., Comment: 5 pages, 4 figures

The interaction of a runaway supermassive black hole (SMBH) with the circumgalactic medium (CGM) can lead to the formation of a wake of shocked gas and young stars behind it. Here we report the serendipitous discovery of an extremely narrow linear feature in HST/ACS images that may be an example of such a wake. The feature extends 62 kpc from the nucleus of a compact star-forming galaxy at z=0.964. Keck LRIS spectra show that the [OIII]/H$\beta$ ratio varies from ~1 to ~10 along the feature, indicating a mixture of star formation and fast shocks. The feature terminates in a bright [OIII] knot with a luminosity of 1.9x10$^{41}$ ergs/s. The stellar continuum colors vary along the feature, and are well-fit by a simple model that has a monotonically increasing age with distance from the tip. The line ratios, colors, and the overall morphology are consistent with an ejected SMBH moving through the CGM at high speed while triggering star formation. The best-fit time since ejection is ~39 Myr and the implied velocity is v~1600 km/s. The feature is not perfectly straight in the HST images, and we show that the amplitude of the observed spatial variations is consistent with the runaway SMBH interpretation. Opposite the primary wake is a fainter and shorter feature, marginally detected in [OIII] and the rest-frame far-ultraviolet. This feature may be shocked gas behind a binary SMBH that was ejected at the same time as the SMBH that produced the primary wake., Comment: Accepted for publication in ApJ Letters. The key data are in Figure 1: a really odd thin streak in HST images, with a complex emission line spectrum. Figure 7 is an illustration of our proposed interpretation

With a constant increase of learned parameters, modern neural language models become increasingly more powerful. Yet, explaining these complex model's behavior remains a widely unsolved problem. In this paper, we discuss the role interactive visualization can play in explaining NLP models (XNLP). We motivate the use of visualization in relation to target users and common NLP pipelines. We also present several use cases to provide concrete examples on XNLP with visualization. Finally, we point out an extensive list of research opportunities in this field.

While deep reinforcement learning (RL) agents have showcased strong results across many domains, a major concern is their inherent opaqueness and the safety of such systems in real-world use cases. To overcome these issues, we need agents that can quantify their uncertainty and detect out-of-distribution (OOD) states. Existing uncertainty estimation techniques, like Monte-Carlo Dropout or Deep Ensembles, have not seen widespread adoption in on-policy deep RL. We posit that this is due to two reasons: concepts like uncertainty and OOD states are not well defined compared to supervised learning, especially for on-policy RL methods. Secondly, available implementations and comparative studies for uncertainty estimation methods in RL have been limited. To overcome the first gap, we propose definitions of uncertainty and OOD for Actor-Critic RL algorithms, namely, proximal policy optimization (PPO), and present possible applicable measures. In particular, we discuss the concepts of value and policy uncertainty. The second point is addressed by implementing different uncertainty estimation methods and comparing them across a number of environments. The OOD detection performance is evaluated via a custom evaluation benchmark of in-distribution (ID) and OOD states for various RL environments. We identify a trade-off between reward and OOD detection performance. To overcome this, we formulate a Pareto optimization problem in which we simultaneously optimize for reward and OOD detection performance. We show experimentally that the recently proposed method of Masksembles strikes a favourable balance among the survey methods, enabling high-quality uncertainty estimation and OOD detection while matching the performance of original RL agents.

Many data analysis problems rely on dynamic networks, such as social or communication network analyses. Providing a scalable overview of long sequences of such dynamic networks remains challenging due to the underlying large-scale data containing elusive topological changes. We propose two complementary pixel-based visualizations, which reflect occurrences of selected sub-networks (motifs) and provide a time-scalable overview of dynamic networks: a network-level census (motif significance profiles) linked with a node-level sub-network metric (graphlet degree vectors) views to reveal structural changes, trends, states, and outliers. The network census captures significantly occurring motifs compared to their expected occurrences in random networks and exposes structural changes in a dynamic network. The sub-network metrics display the local topological neighborhood of a node in a single network belonging to the dynamic network. The linked pixel-based visualizations allow exploring motifs in different-sized networks to analyze the changing structures within and across dynamic networks, for instance, to visually analyze the shape and rate of changes in the network topology. We describe the identification of visual patterns, also considering different reordering strategies to emphasize visual patterns. We demonstrate the approach's usefulness by a use case analysis based on real-world large-scale dynamic networks, such as the evolving social networks of Reddit or Facebook., Comment: 10 pages, 5 figures

Robustness to adversarial perturbations has been explored in many areas of computer vision. This robustness is particularly relevant in vision-based reinforcement learning, as the actions of autonomous agents might be safety-critic or impactful in the real world. We investigate the susceptibility of vision-based reinforcement learning agents to gradient-based adversarial attacks and evaluate a potential defense. We observe that Bottleneck Attention Modules (BAM) included in CNN architectures can act as potential tools to increase robustness against adversarial attacks. We show how learned attention maps can be used to recover activations of a convolutional layer by restricting the spatial activations to salient regions. Across a number of RL environments, BAM-enhanced architectures show increased robustness during inference. Finally, we discuss potential future research directions., Comment: 5 pages, 2 figures, 3 tables

The Navier-Stokes-Korteweg (NSK) system is a classical diffuse interface model which is based on van der Waals theory of capillarity. Diffuse interface methods have gained much interest to model two-phase flow in porous media. However, for the numerical solution of the NSK equations two major challenges have to be faced. First, an extended numerical stencil is required due to a third-order term in the linear momentum and the total energy equations. In addition, the dispersive contribution in the linear momentum equations prevents the straightforward use of contact angle boundary conditions. Secondly, any real gas equation of state is based on a non-convex Helmholtz free energy potential which may cause the eigenvalues of the Jacobian of the first-order fluxes to become imaginary numbers inside the spinodal region. In this work, a thermodynamically consistent relaxation model is presented which is used to approximate the NSK equations. The model is complimented by thermodynamically consistent non-equilibrium boundary conditions which take contact angle effects into account. Due to the relaxation approach, the contribution of the Korteweg tensor in the linear momentum and total energy equations can be reduced to second-order terms which enables a straightforward implementation of contact angle boundary conditions in a numerical scheme. Moreover, the definition of a modified pressure function enables to formulate first-order fluxes which remain strictly hyperbolic in the entire spinodal region. The present work is a generalization of a previously presented parabolic relaxation model for the isothermal NSK equations.

The Laser Interferometer Space Antenna (LISA) will unveil the non-transient gravitational wave sky full of inspiralling stellar-mass compact binaries within the Local Universe. The Large Magellanic Cloud (LMC) is expected to be prominent on the LISA sky due to its proximity and its large population of double white dwarfs (DWD). Here we present the first dedicated study of the LMC with gravitational wave sources. We assemble three LMC models based on: (1) the density distribution and star formation history from optical wavelength observations, (2) a detailed hydrodynamic simulation, and (3) combining the two. Our models yield a hundred to several hundred detectable DWDs: indeed, the LMC will be a resolved galaxy in the LISA sky. Importantly, amongst these we forecast a few tens to a hundred double degenerate supernovae type Ia progenitors, a class of binaries which have never been unambiguously observed. The range in the number of detections is primarily due to differences in the LMC total stellar mass and recent star formation in our models. Our results suggest that the total number, periods, and chirp masses of LISA sources may provide independent constraints on both LMC stellar mass and recent star formation by comparing LISA observations with the models, although such constraints will be highly model-dependent. Our publicly available model populations may be used in future studies of the LMC, including its structure and contribution to LISA confusion noise., Comment: Accepted to MNRAS. Adapted from M. A. Keim's MSc Thesis. Catalogues available at https://doi.org/10.5281/zenodo.6918083