Interpersonal communication in data science can yield sought-after insights, but presentation environments are often not conducive for live analysis, forcing the process to move offline. Through a formative survey with 16 participants, we identified both technical (e.g., complexity of tools) and psychological (e.g., pressure of programming during presentation) factors constraining live data analysis. To enable live analysis, we present SlideSAVR, a data-driven presentation assistant that leverages sketching and voice inputs in live discussion to support collaborative data analysis during presentations. Powered by an agentic framework that flexibly defines augmentation rules, updates slide content dynamically to match the live context, and automates backend computations, SlideSAVR enables fluid audience-presenter interaction and reduces the need for offline reanalysis and follow-up communication. We demonstrate SlideSAVR’s ability to support a range of tasks through nine representative use cases. We further evaluate the system’s accuracy and computation time across different settings, showing that SlideSAVR can reliably perform diverse tasks when provided with both sketch and voice inputs.

What makes a Gantt chart? This question proved unexpectedly difficult to answer when we set out to build a design space for Gantt charts. Existing definitions, each shaped by their respective research goals, made different scope choices that we could not directly reconcile. We reasoned about what should and should not count as a Gantt chart, developing concepts and tools along the way. We distinguish features that are essential to a chart type's identity from those that can vary, and use these distinctions to map how chart types relate through what they share and lack. Applying these ideas to Gantt charts, radar charts, and table cartograms, we produce key insights on what boundary work reveals: definitions diverge for functional reasons, drawing boundaries exposes hidden structure in descriptive vocabulary such as feature entanglements, and scope choices shape how far findings can generalize. We came to understand that there is not a definitive answer, but that working through the question produced a functional definition that guided the design space we originally set out to build. Additionally, we present vocabulary and tools for reasoning about chart type boundaries and surfacing these boundary decisions, alongside a documented Gantt chart design space. Our broader reflection is that scope choices in chart-type-centered research---which determine what design spaces include, what grammars generate, and what perceptual studies measure---are research decisions worth making visible.

Time series forecasting is a fundamental tool with wide ranging applications, yet recent debates question whether complex nonlinear architectures truly outperform simple linear models. Prior claims of dominance of the linear model often stem from benchmarks that lack diverse temporal dynamics and employ biased evaluation protocols. We revisit this debate through TimeSynth, a structured framework that emulates key properties of real world time series,including non-stationarity, periodicity, trends, and phase modulation by creating synthesized signals whose parameters are derived from real-world time series. Evaluating four model families Linear, Multi Layer Perceptrons (MLP), Convolutional Neural Networks (CNNs), and Transformers, we find a systematic bias in linear models: they collapse to simple oscillation regardless of signal complexity. Nonlinear models avoid this collapse and gain clear advantages as signal complexity increases. Notably, Transformers and CNN based models exhibit slightly greater adaptability to complex modulated signals compared to MLPs. Beyond clean forecasting, the framework highlights robustness differences under distribution and noise shifts and removes biases of prior benchmarks by using independent instances for train, test, and validation for each signal family. Collectively, TimeSynth provides a principled foundation for understanding when different forecasting approaches succeed or fail, moving beyond oversimplified claims of model equivalence. 

While medical Vision-Language models (VLMs) achieve strong performance on tasks such as tumor or organ segmentation and diagnosis prediction, their opaque latent representations limit clinical trust and the ability to explain predictions. Interpretability of these multimodal representations are therefore essential for the trustworthy clinical deployment of pretrained medical VLMs. However, current interpretability methods, such as gradient- or attention-based visualizations, are often limited to specific tasks such as classification. Moreover, they do not provide concept-level explanations derived from shared pretrained representations that can be reused across downstream tasks. We introduce MedConcept, a framework that uncovers latent medical concepts in a fully unsupervised manner and grounds them in clinically verifiable textual semantics. MedConcept identifies sparse neuron-level concept activations from pretrained VLM representations and translates them into pseudo-report-style summaries, enabling physician-level inspection of internal model reasoning. To address the lack of quantitative evaluation in concept-based interpretability, we introduce a quantitative semantic verification protocol that leverages an independent pretrained medical LLM as a frozen external evaluator to assess concept alignment with radiology reports. We define three concept scores, Aligned, Unaligned, and Uncertain, to quantify semantic support, contradiction, or ambiguity relative to radiology reports and use them exclusively for post hoc evaluation. These scores provide a quantitative baseline for assessing interpretability in medical VLMs.

Forecasting models for health-signal digital twins must preserve the oscillatory, frequency, phase, and state-transition dynamics of physiological signals, yet the pointwise metrics used to benchmark them cannot detect when these fundamental properties are lost. We show that this blind spot misranks models: across 11 architectures, models with comparable pointwise error diverge by up to 53° in phase accuracy, equivalent to roughly 123 ms for a 1.2 Hz cardiac rhythm and invisible to standard metrics. To enable development of models that escape such failures, we introduce TimeSynth, a controlled benchmarking framework with two reusable components: a physiologically grounded generator producing signals with analytically known ground-truth dynamics from parametric models fitted to real electroencephalography, electrocardiography and photoplethysmogram signals, along with diagnostics quantifying amplitude, frequency, phase, and state-transition fidelity. Linear and full-sequence attention models systematically lose frequency and phase information despite acceptable amplitude error, whereas architectures with localized temporal structure better preserve dynamical fidelity and adapt to observable state transitions; none, however, reliably preserves stochastic switching. Because the dominant determinant of fidelity is architectural, model choice becomes a principled, use-case-driven decision rather than a search for a single winner. TimeSynth thus supplies the controlled preclinical stress test missing before models are coupled to patient data, with a reusable generator and diagnostics for fidelity-aware development.

Dense voxel-level annotation remains a major bottleneck in 3D medical image segmentation. Single-slice propagation methods such as Sli2Vol reduce this burden by propagating one annotated seed slice through a volume using label-free registration. However, axial-only propagation accumulates errors with distance from the seed, especially in surface-distance metrics, because it ignores coronal and sagittal evidence and therefore underuses the 3D information available in CT/MRI volumes. To better leverage volumetric geometry, we study how key training and inference choices affect slice-propagation models, including single-axis versus multi-axis label-free registration, single-seed versus multi-seed propagation, and orthogonal seed configurations. Instead of propagating from a single axial seed, we use three orthogonal seeds---one axial, one coronal, and one sagittal---and fuse their propagated labels with a simple label-free rule. Our results show that the training paradigm has limited impact: an axially trained network applied to off-axis seeds captures nearly all the improvement, while explicit three-axis training adds little. Instead, performance is driven by inference-time seed geometry, especially orthogonality rather than the number of annotated slices, as a budget-matched three-axial control provides no benefit and can even degrade performance. On a multi-organ CT cohort, orthogonal seeding with the axial Sli2Vol backbone improves Dice by 21.9%, Normalized Surface Dice by 25.5%, and reduces Average Hausdorff Distance by 53.5% over the single-axis baseline.

Hyper-differential sensitivity analysis with respect to model discrepancy was recently developed to enable uncertainty quantification for optimization problems. The approach consists of two primary steps: (i) Bayesian calibration of the discrepancy between high- and low-fidelity models, and (ii) propagating the model discrepancy uncertainty through the optimization problem. When high-fidelity model evaluations are limited, as is common in practice, the prior discrepancy distribution plays a crucial role in the uncertainty analysis. However, specification of this prior is challenging due to its mathematical complexity and many hyper-parameters. This article presents a novel approach to specify the prior distribution. Our approach consists of two parts: (1) an algorithmic initialization of the prior hyper-parameters that uses existing data to initialize a hyper-parameter estimate, and (2) a visualization framework to systematically explore properties of the prior and guide tuning of the hyper-parameters to ensure that the prior captures the appropriate range of uncertainty. We provide detailed mathematical analysis and a collection of numerical examples that elucidate properties of the prior that are crucial to ensure uncertainty quantification.

Quantum processors are being integrated into HPC ecosystems as co-processors, where compilation of quantum circuits into hardware-executable form determines both output fidelity and runtime. Current compilers use a fixed pass sequence and ignore the fact that optimal pass selection varies with circuit, hardware, and noise conditions. We present TuniQ, a reinforcement learning-based system that selects compilation passes at each pipeline stage, adapting to circuit, backend, and current noise profile. TuniQ introduces several novel design components like a dual-encoder for stage-aware representation, shaped rewards for cross-stage credit assignment, and dynamic action masking for valid compilation. Evaluated across diverse quantum workloads on multiple IBM Quantum Cloud processors, TuniQ improves fidelity and reduces compilation time over the state-of-the-art IBM Qiskit transpiler, generalizes across backends without retraining, and scales strongly to utility-scale circuits with growing advantage.

Wrangling consumes a disproportionate share of the effort associated with any data project. While a variety of tools support it, relatively little is known about how their differing interface forms shape the way people actually wrangle. We conduct a between-subjects (N=40) observational study of data cleaning tasks performed in tools spanning distinct interface paradigms: Jupyter (notebook), Excel (spreadsheet), ChatGPT (conversational AI), and OpenRefine (visual wranglers). We situate our observations within the Technical Dimensions of Programming Systems framework, which we use as a conceptual scaffold for comparing across interface paradigms. Within the context of our study, the results suggest that tool affordances steer user strategies but do not determine outcomes. There is no consistent advantage of any single tool, nor convergence of results within tools observed across our outcome measures. Instead, we identify trade-offs and connect them with observed practice. For example, a key tension is between data- and abstraction-centered interfaces, where data-centered interfaces encourage opportunistic cleaning rather than systematic, planned transformations found in abstraction-focused tools (but come with a cognitive burden). Tool design, beyond mere functionality, plays a structuring role in how data work unfolds.

Textual data serialization formats, such as JSON or XML, are ubiquitous, supporting tasks like software configuration and data tabularization. Despite their prominence, little is known about their usability. What makes one good or bad? Is there a best one for cognitive efficiency? We explore these questions via a (N=215) crowd work study and a (N=9) semi-structured interview study. We find that format distinctions (like indentation versus curly braces) do not consistently translate into substantial usability differences. While HJSON and YAML performed better than other formats in certain modification tasks, these advantages disappeared in more realistic settings where task complexity was either trivial or highly demanding. Instead, usability appears driven by sociotechnical ecosystems: the tooling, documentation, and community practices surrounding a format matter more than syntax.

Visualizations are central to communicating data, yet blind and low-vision (BLV) people often lack support for understanding chart types---knowledge that is essential for interpreting new visualizations and collaborating with sighted peers. Prior work found that BLV individuals viewed example tactile charts as more helpful than text-only approaches and preferred them for learning advanced chart types, particularly for understanding spatial layouts and shapes. Meanwhile, large language models (LLMs) are increasingly used by BLV individuals for chart explanation and question answering (QA), but have been studied primarily for dataset exploration rather than chart-type learning. Existing LLM-based chart QA also shows that users frequently ask about layout and structure, yet struggle with spatial concepts and misdirect questions when mental models are weak. We investigate how LLMs influence chart-type learning and whether tactile learning improves subsequent LLM-supported exploration. We extend our tactile chart learning tools with an LLM chatbot that provides interactive explanations and supports follow-up questions. In an interview study with 12 BLV participants, we compare two learning formats: (1) a tactile chart, a textual explanation, and an LLM chatbot; and (2) a textual explanation and an LLM chatbot. The learning phase was followed by exploration of an unfamiliar dataset using alt text and an LLM. Thematic analysis shows that tactile templates support BLV participants' formation of chart-type mental models, which scaffolds subsequent LLM-mediated data exploration. Text+LLM explanations without tactile support show weaknesses for spatial-reasoning tasks.

Designing a color palette for data requires designers to balance multiple constraints, including accessibility and aesthetics. Color palette tools support this process through features including direct manipulation, automated palette generation and evaluation, previews, and so on. Despite their prominence, relatively little is known about how these different mechanisms shape design across contexts. We conducted an exploratory think-aloud crowd work study with 40 self-identified designers. Each participant used one of four palette tools selected to span different interaction modalities to complete a series of accessibility- and aesthetics-oriented design tasks. We observed two preliminary patterns. First, tool differences were more pronounced in accessibility-constrained tasks. Second, even when accessibility was not explicitly required, some tools produced more accessibility-friendly palettes and prompted more accessibility-oriented thinking. In this tool genre, then, system design shapes outcomes both via built-in functionality, as well as by directing designers' attention toward particular constraints and design considerations.

Accelerated 3D late gadolinium enhancement (LGE) MRI requires robust reconstruction methods to recover thin atrial structures from undersampled k-space data. While unrolled model-based networks effectively integrate physics-driven data consistency with learned priors, they operate at the acquired resolution and may fail to fully recover high-frequency detail. We propose a hybrid unrolled reconstruction framework in which an Enhanced Deep Super-Resolution (EDSR) network replaces the proximal operator within each iteration of the optimization loop, enabling joint super-resolution enhancement and data consistency enforcement. The model is trained end-to-end on retrospectively undersampled preclinical 3D LGE datasets and compared against compressed sensing, Model-Based Deep Learning (MoDL), and self-guided Deep Image Prior (DIP) baselines. Across acceleration factors, the proposed method consistently improves PSNR and SSIM over standard unrolled reconstruction and better preserves fine cardiac structures, leading to improved LA (left atrium) segmentation performance. These results demonstrate that integrating super-resolution priors directly within model-based reconstruction provides measurable gains in accelerated 3D LGE MRI.

Mean flow (MeanFlow) enables efficient, high-fidelity image generation, yet its single-function evaluation (1-NFE) generation often cannot yield compelling results. We address this issue by introducing RMFlow, an efficient multimodal generative model that integrates a coarse 1-NFE MeanFlow transport with a subsequent tailored noise-injection refinement step. RMFlow approximates the average velocity of the flow path using a neural network trained with a new loss function that balances minimizing the Wasserstein distance between probability paths and maximizing sample likelihood. RMFlow achieves near state-of-the-art results on text-to-image, context-to-molecule, and time-series generation using only 1-NFE, at a computational cost comparable to the baseline MeanFlows. 

Patient-specific finite element models of hip cartilage and labrum (chondrolabral) mechanics have improved the understanding of form-function relationships underpinning hip osteoarthritis. While these models often assume the pelvis and femur are rigid bodies to reduce development and computational time, contact stresses may be overestimated when the pelvis and femur cannot deform. Recent advancements in element formulations, constitutive models, and integration of patient-specific boundary/loading conditions warrant a re-examination of this rigid body assumption. We assessed the influence of material representations for the pelvis and femur on finite element predictions of chondrolabral mechanics during level walking and squatting. Four material representations were evaluated: (1) rigid bodies, (2) deformable-inhomogeneous with CT-derived material properties, (3) deformable-two-material with distinct trabecular and cortical bone materials, and (4) deformable-one-material model. Patient-specific kinematics and joint reaction forces for level walking and squatting were derived from motion capture and musculoskeletal models, respectively. During level walking, there were no significant differences between bone material representations for maximum cartilage contact pressure, contact area, shear stress, or first principal strain. During squatting, rigid and one-material models produced significantly higher maximum contact pressures, while other metrics remained unaffected. Rigid models required significantly less computational runtime and memory than deformable models. Our findings indicate rigid bone assumptions suffice for level walking, improving efficiency in large cohort studies, but deformable bone models are likely warranted for activities that produce larger bone deformations such as squatting. This study provides guidance for selecting bone material representations that balance accuracy and computational efficiency in hip biomechanical analyses.

Subject-specific finite element (FE) modeling enables detailed evaluation of cartilage and labral contact mechanics in the hip; however, the computational demands of this approach limit its use in large cohorts. Soft tissue overlap (STO) modeling is a simplified alternative that estimates strain from geometric overlap between articulating surfaces; however, its predictive accuracy remains unclear. The objective of this study was to evaluate the accuracy of STO models for predicting acetabular cartilage and labral strains during simulated walking by direct comparison with subject-specific FE models. Eight individuals with radiographically normal hips and eight patients with cam-type femoroacetabular impingement syndrome were evaluated using subject-specific anatomy, kinematics, and joint reaction forces. STO-predicted strains were compared with FE-predicted compressive strains using correspondence-based Network Analysis and Bland–Altman analyses at heel-strike and heel-off. No group-dependent differences in agreement were observed, so the results for the two cohorts were pooled. STO models captured the general spatial patterns and locations of peak strain in both cartilage and labrum. However, significant differences between STO and FE predictions were present across 1% to 75% of the gait cycle, particularly during higher magnitudes of joint loading. STO systematically underpredicted strain at higher magnitudes, with disagreement increasing as strain magnitude increased. Discrepancies were most pronounced in the anterosuperior cartilage and labrum near heel-off, where FE models predicted higher, more localized strains. These findings indicate that while STO modeling provides qualitative insight into strain patterns and contact locations, its ability to quantify strain magnitude, particularly under higher loading conditions and in the labrum, is limited.

Color discrimination is a fundamental aspect of visualization as it influences how people interpret visual encodings. Many visualization guidelines are informed by perceptual studies, yet relatively few have been replicated. Acknowledging that the interaction between human perception, visual tasks, and display technology can change over time, we replicate two crowdsourced color discrimination studies conducted 10 years earlier. Specifically, we replicated a visualization-focused color discrimination task (N=144) and a more general perceptual discrimination task (N=394). In both studies, our results reproduced the original perceptual effects. We further use the replication to investigate whether color-related practice influences color discrimination. Specifically, we extended our replication studies by adding questions about participants' engagement with color practices. We then examined whether diverse color-related practices (e.g., artistic hobbies, knowledge of color theory, and cosmetic makeup use) influenced color discrimination. We found no significant difference between participants who reported engaging in color-related practices and those who did not, suggesting that design guidance regarding color discrimination may generalize across viewers regardless of their regular color practice.

Scientists face significant visualization challenges as time-varying datasets grow in speed and volume, often requiring specialized infrastructure and expertise to handle massive datasets. Petascale climate models generated in NASA laboratories require a dedicated group of graphics and media experts and access to high-performance computing resources. Scientists may need to share scientific results with the community iteratively and quickly. However, the time-consuming trial-and-error process incurs significant data transfer overhead and far exceeds the time and resources allocated for typical post-analysis visualization tasks, disrupting the production workflow. Our paper introduces a user-friendly framework for creating 3D animations of petascale, time-varying data on a commodity workstation. Our contributions: (i) Generalized Animation Descriptor (GAD) with a keyframe-based adaptable abstraction for animation, (ii) efficient data access from cloud-hosted repositories to reduce data management overhead, (iii) tailored rendering system, and (iv) an LLM-assisted conversational interface as a scripting module to allow domain scientists with no visualization expertise to create animations of their region of interest. We demonstrate the framework's effectiveness with two case studies: first, by generating animations in which sampling criteria are specified based on prior knowledge, and second, by generating AI-assisted animations in which sampling parameters are derived from natural-language user prompts. In all cases, we use large-scale NASA climate-oceanographic datasets that exceed 1PB in size yet achieve a fast turnaround time of 1 minute to 2 hours. Users can generate a rough draft of the animation within minutes, then seamlessly incorporate as much high-resolution data as needed for the final version.

Conducting a complete computer architecture simulation study is challenging because configuration, execution, and analysis are often encoded implicitly in scripts or directory conventions rather than represented explicitly. As a result, studies are difficult to scale, hard to reproduce, and dependent on custom tooling at every stage. We present ArchSim, which makes the structure of a simulation study explicit. In ArchSim, hardware topologies are described as declarative graphs that automatically generate executable simulation code, eliminating hand-written simulator programs. Stateless runners autonomously claim and execute jobs from a shared experiment store, enabling configuration-benchmark matrices to scale without manual orchestration. Simulation outputs are stored as structured artifacts tied to configurations, benchmarks, and hardware components, enabling systematic result exploration without custom parsers. We evaluate ArchSim on a 12 x 8 = 96-configuration simulation matrix spanning memory-bound, compute-bound, and mixed-intensity GPU workloads. Declarative simulation specifications drive full simulations with a median kernel time error of 0.18% relative to hand-written MGPUSim configurations across 95.8% of configurations. The platform introduces only 1.6 seconds of overhead per simulation, negligible relative to realistic simulation workloads.

Hybrid cooling systems that combine wet and dry cooling have drawn growing attention, as they can leverage ambient air temperature to enable broader regional deployment while mitigating water consumption. However, hybrid cooling remains insufficiently explored because the water associated with power generation is often overlooked. Moreover, distributing water is also challenging when water is scarce. In this paper, we present WaterSplit, a water-aware optimization framework that minimizes the water footprint to enable sustainable datacenter operations with hybrid cooling under limited water budgets.