Marco Cavallo

Kevin Miao, Harsh Agrawal, Qihang Zhang et al. · apple-ml, gatech

Generating high-quality 3D content requires models capable of learning robust distributions of complex scenes and the real-world objects within them. Recent Gaussian-based 3D reconstruction techniques have achieved impressive results in recovering high-fidelity 3D assets from sparse input images by predicting 3D Gaussians in a feed-forward manner. However, these techniques often lack the extensive priors and expressiveness offered by Diffusion Models. On the other hand, 2D Diffusion Models, which have been successfully applied to denoise multiview images, show potential for generating a wide range of photorealistic 3D outputs but still fall short on explicit 3D priors and consistency. In this work, we aim to bridge these two approaches by introducing DSplats, a novel method that directly denoises multiview images using Gaussian Splat-based Reconstructors to produce a diverse array of realistic 3D assets. To harness the extensive priors of 2D Diffusion Models, we incorporate a pretrained Latent Diffusion Model into the reconstructor backbone to predict a set of 3D Gaussians. Additionally, the explicit 3D representation embedded in the denoising network provides a strong inductive bias, ensuring geometrically consistent novel view generation. Our qualitative and quantitative experiments demonstrate that DSplats not only produces high-quality, spatially consistent outputs, but also sets a new standard in single-image to 3D reconstruction. When evaluated on the Google Scanned Objects dataset, DSplats achieves a PSNR of 20.38, an SSIM of 0.842, and an LPIPS of 0.109.

Marco Cavallo

While the interpretation of high-dimensional datasets has become a necessity in most industries, and is supported by continuous advances in data science and machine learning, the spatial visualization of higher-dimensional geometry has mostly remained a niche research topic for mathematicians and physicists. Intermittent contributions to this field date back more than a century, and have had a non-negligible influence on contemporary art and philosophy. However, most contributions have focused on the understanding of specific mathematical shapes, with few concrete applications. In this work, we attempt to revive the community's interest in visualizing higher dimensional geometry by shifting the focus from the visualization of abstract shapes to the design of a broader hyper-universe concept, wherein 3D and 4D objects can coexist and interact with each other. Specifically, we discuss the content definition, authoring patterns, and technical implementations associated with the process of extending standard 3D applications as to support 4D mechanics. We operationalize our ideas through the introduction of a new hybrid 3D/4D videogame called Across Dimensions, which we developed in Unity3D through the integration of our own 4D plugin.

Marco Cavallo, Mishal Dholakia, Matous Havlena et al.

In the past few years, augmented reality (AR) and virtual reality (VR) technologies have experienced terrific improvements in both accessibility and hardware capabilities, encouraging the application of these devices across various domains. While researchers have demonstrated the possible advantages of AR and VR for certain data science tasks, it is still unclear how these technologies would perform in the context of exploratory data analysis (EDA) at large. In particular, we believe it is important to better understand which level of immersion EDA would concretely benefit from, and to quantify the contribution of AR and VR with respect to standard analysis workflows. In this work, we leverage a Dataspace reconfigurable hybrid reality environment to study how data scientists might perform EDA in a co-located, collaborative context. Specifically, we propose the design and implementation of Immersive Insights, a hybrid analytics system combining high-resolution displays, table projections, and augmented reality (AR) visualizations of the data. We conducted a two-part user study with twelve data scientists, in which we evaluated how different levels of data immersion affect the EDA process and compared the performance of Immersive Insights with a state-of-the-art, non-immersive data analysis system.

Marco Cavallo, Mishal Dholakia, Matous Havlena et al.

Immersive environments have gradually become standard for visualizing and analyzing large or complex datasets that would otherwise be cumbersome, if not impossible, to explore through smaller scale computing devices. However, this type of workspace often proves to possess limitations in terms of interaction, flexibility, cost and scalability. In this paper we introduce a novel immersive environment called Dataspace, which features a new combination of heterogeneous technologies and methods of interaction towards creating a better team workspace. Dataspace provides 15 high-resolution displays that can be dynamically reconfigured in space through robotic arms, a central table where information can be projected, and a unique integration with augmented reality (AR) and virtual reality (VR) headsets and other mobile devices. In particular, we contribute novel interaction methodologies to couple the physical environment with AR and VR technologies, enabling visualization of complex types of data and mitigating the scalability issues of existing immersive environments. We demonstrate through four use cases how this environment can be effectively used across different domains and reconfigured based on user requirements. Finally, we compare Dataspace with existing technologies, summarizing the trade-offs that should be considered when attempting to build better collaborative workspaces for the future.

Marco Cavallo, Çağatay Demiralp

Dimensionality reduction is a common method for analyzing and visualizing high-dimensional data. However, reasoning dynamically about the results of a dimensionality reduction is difficult. Dimensionality-reduction algorithms use complex optimizations to reduce the number of dimensions of a dataset, but these new dimensions often lack a clear relation to the initial data dimensions, thus making them difficult to interpret. Here we propose a visual interaction framework to improve dimensionality-reduction based exploratory data analysis. We introduce two interaction techniques, forward projection and backward projection, for dynamically reasoning about dimensionally reduced data. We also contribute two visualization techniques, prolines and feasibility maps, to facilitate the effective use of the proposed interactions. We apply our framework to PCA and autoencoder-based dimensionality reductions. Through data-exploration examples, we demonstrate how our visual interactions can improve the use of dimensionality reduction in exploratory data analysis.

Marco Cavallo, Angus G. Forbes

Despite advances in augmented reality (AR), the process of creating meaningful experiences with this technology is still extremely challenging. Due to different tracking implementations and hardware constraints, developing AR applications either requires low-level programming skills, or is done through specific authoring tools that largely sacrifice the possibility of customizing the AR experience. Existing development workflows also do not support previewing or simulating the AR experience, requiring a lengthy process of trial and error by which content creators deploy and physically test applications in each iteration. To mitigate these limitations, we propose CAVE-AR, a novel virtual reality system for authoring, simulating and debugging custom AR experiences. Available both as a standalone or a plug-in tool, CAVE-AR is based on the concept of representing in the same global reference system both in AR content and tracking information, mixing geographical information, architectural features, and sensor data to simulate the context of an AR experience. Thanks to its novel abstraction of existing tracking technologies, CAVE-AR operates independently of users' devices, and integrates with existing programming tools to provide maximum flexibility. Our VR application provides designers with ways to create and modify an AR application, even while others are in the midst of using it. CAVE-AR further allows the designer to track how users are behaving, preview what they are currently seeing, and interact with them through several different channels. To illustrate our proposed development workflow and demonstrate the advantages of our authoring system, we introduce two CAVEAR use cases in which an augmented reality application is created and tested. We compare the CAVE-AR workflow to traditional development methods and demonstrate the importance of simulation and live application debugging.

Marco Cavallo, Çağatay Demiralp

With the rapid commoditization of wearable sensors, detecting human movements from sensor datasets has become increasingly common over a wide range of applications. To detect activities, data scientists iteratively experiment with different classifiers before deciding which model to deploy. Effective reasoning about and comparison of alternative classifiers are crucial in successful model development. This is, however, inherently difficult in developing classifiers for sensor data, where the intricacy of long temporal sequences, high prediction frequency, and imprecise labeling make standard evaluation methods relatively ineffective and even misleading. We introduce Track Xplorer, an interactive visualization system to query, analyze, and compare the predictions of sensor-data classifiers. Track Xplorer enables users to interactively explore and compare the results of different classifiers, and assess their accuracy with respect to the ground-truth labels and video. Through integration with a version control system, Track Xplorer supports tracking of models and their parameters without additional workload on model developers. Track Xplorer also contributes an extensible algebra over track representations to filter, compose, and compare classification outputs, enabling users to reason effectively about classifier performance. We apply Track Xplorer in a collaborative project to develop classifiers to detect movements from multisensor data gathered from Parkinson's disease patients. We demonstrate how Track Xplorer helps identify early on possible systemic data errors, effectively track and compare the results of different classifiers, and reason about and pinpoint the causes of misclassifications.

Marco Cavallo, Çağatay Demiralp

Data clustering is a common unsupervised learning method frequently used in exploratory data analysis. However, identifying relevant structures in unlabeled, high-dimensional data is nontrivial, requiring iterative experimentation with clustering parameters as well as data features and instances. The number of possible clusterings for a typical dataset is vast, and navigating in this vast space is also challenging. The absence of ground-truth labels makes it impossible to define an optimal solution, thus requiring user judgment to establish what can be considered a satisfiable clustering result. Data scientists need adequate interactive tools to effectively explore and navigate the large clustering space so as to improve the effectiveness of exploratory clustering analysis. We introduce \textit{Clustrophile~2}, a new interactive tool for guided clustering analysis. \textit{Clustrophile~2} guides users in clustering-based exploratory analysis, adapts user feedback to improve user guidance, facilitates the interpretation of clusters, and helps quickly reason about differences between clusterings. To this end, \textit{Clustrophile~2} contributes a novel feature, the Clustering Tour, to help users choose clustering parameters and assess the quality of different clustering results in relation to current analysis goals and user expectations. We evaluate \textit{Clustrophile~2} through a user study with 12 data scientists, who used our tool to explore and interpret sub-cohorts in a dataset of Parkinson's disease patients. Results suggest that \textit{Clustrophile~2} improves the speed and effectiveness of exploratory clustering analysis for both experts and non-experts.

Marco Cavallo, Çağatay Demiralp

Detecting motor activities from sensor datasets is becoming increasingly common in a wide range of applications with the rapid commoditization of wearable sensors. To detect activities, data scientists iteratively experiment with different classifiers before deciding on a single model. Evaluating, comparing, and reasoning about prediction results of alternative classifiers is a crucial step in the process of iterative model development. However, standard aggregate performance metrics (such as accuracy score) and textual display of individual event sequences have limited granularity and scalability to effectively perform this critical step. To ameliorate these limitations, we introduce Track Xplorer, an interactive visualization system to query, analyze and compare the classification output of activity detection in multi-sensor data. Track Xplorer visualizes the results of different classifiers as well as the ground truth labels and the video of activities as temporally-aligned linear tracks. Through coordinated track visualizations, Track Xplorer enables users to interactively explore and compare the results of different classifiers, assess their accuracy with respect to the ground truth labels and video. Users can brush arbitrary regions of any classifier track, zoom in and out with ease, and playback the corresponding video segment to contextualize the performance of the classifier within the selected region. Track Xplorer also contributes an algebra over track representations to filter, compose, and compare classification outputs, enabling users to effectively reason about the performance of classifiers. We demonstrate how our tool helps data scientists debug misclassifications and improve the prediction performance in developing activity classifiers for real-world, multi-sensor data gathered from Parkinson's patients.

Marco Cavallo, Çağatay Demiralp

Dimensionality reduction is a common method for analyzing and visualizing high-dimensional data across domains. Dimensionality-reduction algorithms involve complex optimizations and the reduced dimensions computed by these algorithms generally lack clear relation to the initial data dimensions. Therefore, interpreting and reasoning about dimensionality reductions can be difficult. In this work, we introduce two interaction techniques, \textit{forward projection} and \textit{backward projection}, for reasoning dynamically about scatter plots of dimensionally reduced data. We also contribute two related visualization techniques, \textit{prolines} and \textit{feasibility map} to facilitate and enrich the effective use of the proposed interactions, which we integrate in a new tool called \textit{Praxis}. To evaluate our techniques, we first analyze their time and accuracy performance across varying sample and dimension sizes. We then conduct a user study in which twelve data scientists use \textit{Praxis} so as to assess the usefulness of the techniques in performing exploratory data analysis tasks. Results suggest that our visual interactions are intuitive and effective for exploring dimensionality reductions and generating hypotheses about the underlying data.