Search

We give a definitive characterization of the instability of the pressureless ($p=0$) critical ($k=0$) Friedmann spacetime to smooth radial perturbations. We use this to characterize the global accelerations away from $k\leq0$ Friedmann spacetimes induced by the instability in the underdense case. The analysis begins by incorporating the Friedmann spacetimes into a mathematical analysis of smooth spherically symmetric solutions of the Einstein field equations expressed in self-similar coordinates $(t,\xi)$ with $\xi=\frac{r}{t}<1$, conceived to realize the critical Friedmann spacetime as an unstable saddle rest point $SM$. We identify a new maximal asymptotically stable family $\mathcal{F}$ of smooth outwardly expanding solutions which globally characterize the evolution of underdense perturbations. Solutions in $\mathcal{F}$ align with a $k<0$ Friedmann spacetime at early times, generically introduce accelerations away from $k<0$ Friedmann spacetimes at intermediate times and then decay back to the same $k<0$ Friedmann spacetime as $t\to\infty$ uniformly at each fixed radius $r>0$. We propose $\mathcal{F}$ as the maximal asymptotically stable family of solutions into which generic underdense perturbations of the unstable critical Friedmann spacetime will evolve and naturally admit accelerations away from Friedmann spacetimes within the dynamics of solutions of Einstein's original field equations, that is, without recourse to a cosmological constant or dark energy.

We consider weighted automata over words and over trees where the weight algebras are strong bimonoids, i.e., semirings which may lack distributivity. It is well known that, for each such weighted automaton, its run semantics and its initial algebra semantics can be different, due to the presence of nondeterminism and the absence of distributivity. Here we investigate the question under which conditions on the strong bimonoid the support of the run semantics equals the support of the initial algebra semantics. We prove a characterization of this equality in terms of strongly zero-sum-free strong bimonoids (for weighted automata over words) and in terms of bi-strongly zero-sum-free strong bimonoids (for weighted automata over trees). We also consider shortly the images of the two semantics functions.

In this paper we present a new interpretation of the Lions derivative as the Radon-Nikodym derivative of a vector measure, which provides a canonical extension of the Lions derivative for functions taking values in infinite dimensional Banach spaces. This is of particular relevance for the analysis of Hilbert space valued Mean-Field equations. As an illustration we derive a mild Ito-formula for Mean-Field stochastic partial differential equations (SPDEs), which provides the basis for a higher order Taylor expansion and higher order numerical schemes.

Diffusion models (DMs) have emerged as powerful foundation models for a variety of tasks, with a large focus in synthetic image generation. However, their requirement of large annotated datasets for training limits their applicability in medical imaging, where datasets are typically smaller and sparsely annotated. We introduce DiNO-Diffusion, a self-supervised method for training latent diffusion models (LDMs) that conditions the generation process on image embeddings extracted from DiNO. By eliminating the reliance on annotations, our training leverages over 868k unlabelled images from public chest X-Ray (CXR) datasets. Despite being self-supervised, DiNO-Diffusion shows comprehensive manifold coverage, with FID scores as low as 4.7, and emerging properties when evaluated in downstream tasks. It can be used to generate semantically-diverse synthetic datasets even from small data pools, demonstrating up to 20% AUC increase in classification performance when used for data augmentation. Images were generated with different sampling strategies over the DiNO embedding manifold and using real images as a starting point. Results suggest, DiNO-Diffusion could facilitate the creation of large datasets for flexible training of downstream AI models from limited amount of real data, while also holding potential for privacy preservation. Additionally, DiNO-Diffusion demonstrates zero-shot segmentation performance of up to 84.4% Dice score when evaluating lung lobe segmentation. This evidences good CXR image-anatomy alignment, akin to segmenting using textual descriptors on vanilla DMs. Finally, DiNO-Diffusion can be easily adapted to other medical imaging modalities or state-of-the-art diffusion models, opening the door for large-scale, multi-domain image generation pipelines for medical imaging., Comment: 12 pages, 5 figures

Data is critical to advancing AI technologies, yet its quality and documentation remain significant challenges, leading to adverse downstream effects (e.g., potential biases) in AI applications. This paper addresses these issues by introducing Croissant-RAI, a machine-readable metadata format designed to enhance the discoverability, interoperability, and trustworthiness of AI datasets. Croissant-RAI extends the Croissant metadata format and builds upon existing responsible AI (RAI) documentation frameworks, offering a standardized set of attributes and practices to facilitate community-wide adoption. Leveraging established web-publishing practices, such as Schema.org, Croissant-RAI enables dataset users to easily find and utilize RAI metadata regardless of the platform on which the datasets are published. Furthermore, it is seamlessly integrated into major data search engines, repositories, and machine learning frameworks, streamlining the reading and writing of responsible AI metadata within practitioners' existing workflows. Croissant-RAI was developed through a community-led effort. It has been designed to be adaptable to evolving documentation requirements and is supported by a Python library and a visual editor., Comment: 10 pages, appendix

We consider the images of the initial algebra semantics of weighted tree automata over strong bimonoids (hence also over semirings). These images are subsets of the carrier set of the underlying strong bimonoid. We consider locally finite, weakly locally finite, and bi-locally finite strong bimonoids. We show that there exists a strong bimonoid which is weakly locally finite and not locally finite. We also show that if the ranked alphabet contains a binary symbol, then for any finitely generated strong bimonoid, weighted tree automata can generate, via their initial algebra semantics, all elements of the strong bimonoid. As a consequence of these results, for weakly locally finite strong bimonoids which are not locally finite, weighted tree automata can generate infinite images provided that the input ranked alphabet contains at least one binary symbol. This is in sharp contrast to the setting of weighted string automata, where each such image is known to be finite. As a further consequence, for any finitely generated semiring, there exists a weighted tree automaton which generates, via its run semantics, all elements of the semiring., Comment: 20 pages, 2 figures. arXiv admin note: text overlap with arXiv:2212.05529

While content-based image retrieval (CBIR) has been extensively studied in natural image retrieval, its application to medical images presents ongoing challenges, primarily due to the 3D nature of medical images. Recent studies have shown the potential use of pre-trained vision embeddings for CBIR in the context of radiology image retrieval. However, a benchmark for the retrieval of 3D volumetric medical images is still lacking, hindering the ability to objectively evaluate and compare the efficiency of proposed CBIR approaches in medical imaging. In this study, we extend previous work and establish a benchmark for region-based and localized multi-organ retrieval using the TotalSegmentator dataset (TS) with detailed multi-organ annotations. We benchmark embeddings derived from pre-trained supervised models on medical images against embeddings derived from pre-trained unsupervised models on non-medical images for 29 coarse and 104 detailed anatomical structures in volume and region levels. For volumetric image retrieval, we adopt a late interaction re-ranking method inspired by text matching. We compare it against the original method proposed for volume and region retrieval and achieve a retrieval recall of 1.0 for diverse anatomical regions with a wide size range. The findings and methodologies presented in this paper provide insights and benchmarks for further development and evaluation of CBIR approaches in the context of medical imaging., Comment: 34 pages, 12 Figures, 22 Tables

We propose a novel computational approach to automatically analyze the physical process behind printing of early modern letterpress books via clustering the running titles found at the top of their pages. Specifically, we design and compare custom neural and feature-based kernels for computing pairwise visual similarity of a scanned document's running titles and cluster the titles in order to track any deviations from the expected pattern of a book's printing. Unlike body text which must be reset for every page, the running titles are one of the static type elements in a skeleton forme i.e. the frame used to print each side of a sheet of paper, and were often re-used during a book's printing. To evaluate the effectiveness of our approach, we manually annotate the running title clusters on about 1600 pages across 8 early modern books of varying size and formats. Our method can detect potential deviation from the expected patterns of such skeleton formes, which helps bibliographers understand the phenomena associated with a text's transmission, such as censorship. We also validate our results against a manual bibliographic analysis of a counterfeit early edition of Thomas Hobbes' Leviathan (1651)., Comment: Accepted at ICDAR 2024; updated Acknowledgments in v2

In this study, we assess the usability of interactive personal assistants (IPAs), such as Amazon Alexa, in a simulated kitchen smart home environment, with deaf and hard of hearing users. Participants engage in activities in a way that causes their hands to get dirty. With these dirty hands, they are tasked with two different input methods for IPAs: American Sign Language (ASL) in a Wizard-of-Oz design, and smart home apps with a touchscreen. Usability ratings show that participants significantly preferred ASL over touch-based apps with dirty hands, although not to a larger extent than in comparable previous work with clean hands. Participants also expressed significant enthusiasm for ASL-based IPA interaction in Netpromoter scores and in questions about their overall preferences. Preliminary observations further suggest that having dirty hands may affect the way people sign, which may pose challenges for building IPAs that natively support sign language input., Comment: To appear in Extended Abstracts of the CHI Conference on Human Factors in Computing Systems, CHI EA 2024, May 11-16, 2024, Honolulu, HI, USA. ACM, New York, NY, USA, 9 pages. https://doi.org/10.1145/3613905.3651075

With the recent advancements in intelligent personal assistants (IPAs), their popularity is rapidly increasing when it comes to utilizing Automatic Speech Recognition within households. In this study, we used a Wizard-of-Oz methodology to evaluate and compare the usability of American Sign Language (ASL), Tap to Alexa, and smart home apps among 23 deaf participants within a limited-domain smart home environment. Results indicate a slight usability preference for ASL. Linguistic analysis of the participants' signing reveals a diverse range of expressions and vocabulary as they interacted with IPAs in the context of a restricted-domain application. On average, deaf participants exhibited a vocabulary of 47 +/- 17 signs with an additional 10 +/- 7 fingerspelled words, for a total of 246 different signs and 93 different fingerspelled words across all participants. We discuss the implications for the design of limited-vocabulary applications as a stepping-stone toward general-purpose ASL recognition in the future., Comment: To appear in Proceedings of the Conference on Human Factors in Computing Systems CHI 24, May 11-16, Honolulu, HI, USA, 15 pages. https://doi.org/10.1145/3613904.3642094

This paper presents a mixed methods study on how deaf, hard of hearing and hearing viewers perceive live TV caption quality with captioned video stimuli designed to mirror TV captioning experiences. To assess caption quality, we used four commonly-used quality metrics focusing on accuracy: word error rate, weighted word error rate, automated caption evaluation (ACE), and its successor ACE2. We calculated the correlation between the four quality metrics and viewer ratings for subjective quality and found that the correlation was weak, revealing that other factors besides accuracy affect user ratings. Additionally, even high-quality captions are perceived to have problems, despite controlling for confounding factors. Qualitative analysis of viewer comments revealed three major factors affecting their experience: Errors within captions, difficulty in following captions, and caption appearance. The findings raise questions as to how objective caption quality metrics can be reconciled with the user experience across a diverse spectrum of viewers., Comment: To appear in Proceedings of the Conference on Human Factors in Computing Systems CHI 24, May 11-16, Honolulu, HI, USA, 16 pages. https://doi.org/10.1145/3613904.3641988

Data is a critical resource for machine learning (ML), yet working with data remains a key friction point. This paper introduces Croissant, a metadata format for datasets that creates a shared representation across ML tools, frameworks, and platforms. Croissant makes datasets more discoverable, portable, and interoperable, thereby addressing significant challenges in ML data management. Croissant is already supported by several popular dataset repositories, spanning hundreds of thousands of datasets, enabling easy loading into the most commonly-used ML frameworks, regardless of where the data is stored. Our initial evaluation by human raters shows that Croissant metadata is readable, understandable, complete, yet concise., Comment: Published at the NeurIPS 2024 Datasets and Benchmark Track. A shorter version appeared earlier in Proceedings of ACM SIGMOD/PODS'24 Data Management for End-to-End Machine Learning (DEEM) Workshop https://dl.acm.org/doi/10.1145/3650203.3663326

A wide range of imaging techniques and data formats available for medical images make accurate retrieval from image databases challenging. Efficient retrieval systems are crucial in advancing medical research, enabling large-scale studies and innovative diagnostic tools. Thus, addressing the challenges of medical image retrieval is essential for the continued enhancement of healthcare and research. In this study, we evaluated the feasibility of employing four state-of-the-art pretrained models for medical image retrieval at modality, body region, and organ levels and compared the results of two similarity indexing approaches. Since the employed networks take 2D images, we analyzed the impacts of weighting and sampling strategies to incorporate 3D information during retrieval of 3D volumes. We showed that medical image retrieval is feasible using pretrained networks without any additional training or fine-tuning steps. Using pretrained embeddings, we achieved a recall of 1 for various tasks at modality, body region, and organ level., Comment: 8 pages, 3 figures, 4 tables

Nonlinear optical effects are used to generate coherent light at wavelengths difficult to reach with lasers. Materials periodically poled or nanostructured in the nonlinear susceptibility in three spatial directions are called 3D nonlinear photonic crystals (NPhCs). They enable enhanced nonlinear optical conversion efficiencies, emission control, and simultaneous generation of nonlinear wavelengths. The chemical inertness of efficient second-order nonlinear materials ($\chi^{(2)}$) prohibited their nanofabrication until 2018. The current method is restricted to top-down laser-based techniques limiting the periodicity along z-axis to 10 um. We demonstrate the first bottom-up fabricated 3D NPhC in sol-gel derived barium titanate by soft-nanoimprint lithography: a woodpile with eight layers and periodicities of 1 um (xy-plane) and 300 nm (z-plane). The surface areas exceed $5.3\cdot 10^4$ um^2, which is two orders of magnitude larger than the state-of-the-art. This study is expected to initiate bottom-up fabrication of 3D NPhCs with a supremely strong and versatile nonlinear response., Comment: 11 pages, 4 figures

The quest of a nonlinear optical material that can be easily nanostructured over a large surface area is still ongoing. Here, we demonstrate a nanoimprinted nonlinear barium titanate 2D nanohole array that shows optical properties of a 2D photonic crystal and metasurface, depending on the direction of the optical axis. The challenge of nanostructuring the inert metal-oxide is resolved by direct soft nanoimprint lithography with sol-gel derived barium titanate enabling critical dimensions of 120 nm with aspect ratios of 5. The nanohole array exhibits a photonic bandgap in the infrared range when probed along the slab axis while lattice resonant states are observed in out-of-plane transmission configuration. The enhanced light-matter interaction from the resonant structure enables to increase the second-harmonic generation in the near-UV by a factor of 18 illustrating the potential in the flexible fabrication technique for barium titanate photonic devices.

In this paper, we develop machine learning techniques to identify unknown printers in early modern (c.~1500--1800) English printed books. Specifically, we focus on matching uniquely damaged character type-imprints in anonymously printed books to works with known printers in order to provide evidence of their origins. Until now, this work has been limited to manual investigations by analytical bibliographers. We present a Contrastive Attention-based Metric Learning approach to identify similar damage across character image pairs, which is sensitive to very subtle differences in glyph shapes, yet robust to various confounding sources of noise associated with digitized historical books. To overcome the scarce amount of supervised data, we design a random data synthesis procedure that aims to simulate bends, fractures, and inking variations induced by the early printing process. Our method successfully improves downstream damaged type-imprint matching among printed works from this period, as validated by in-domain human experts. The results of our approach on two important philosophical works from the Early Modern period demonstrate potential to extend the extant historical research about the origins and content of these books., Comment: Proceedings of AAAI 2023

The continuous introduction of renewable electricity and increased consumption through electrification of the transport and heating sector challenges grid stability. This study investigates load shifting through demand side management as a solution. We present a four-month experimental study of a low-complexity, hierarchical Model Predictive Control approach for demand side management in a near-zero emission occupied single-family house in Denmark. The control algorithm uses a price signal, weather forecast, a single-zone building model, and a non-linear heat pump efficiency model to generate a space-heating schedule. The weather-compensated, commercial heat pump is made to act smart grid-ready through outdoor temperature input override to enable heat boosting and forced stops to accommodate the heating schedule. The cost reduction from the controller ranged from 2-33% depending on the chosen comfort level. The experiment demonstrates that load shifting is feasible and cost-effective, even without energy storage, and that the current price scheme provides an incentive for Danish end-consumers to shift heating loads. However, issues related to controlling the heat pump through input-manipulation were identified, and the authors propose a more promising path forward involving coordination with manufacturers and regulators to make commercial heat pumps truly smart grid-ready.

In this paper we present a method to approximate optimal feedback controls for stochastic reaction-diffusion equations. We derive two approximation results providing the theoretical foundation of our approach and allowing for explicit error estimates. The approximation of optimal feedback controls by neural networks is discussed as an explicit application of our method. Finally we provide a numerical examples to illustrate our findings.

We present a numerical algorithm that allows the approximation of optimal controls for stochastic reaction-diffusion equations with additive noise by first reducing the problem to controls of feedback form and then approximating the feedback function using finitely based approximations. Using structural assumptions on the finitely based approximations, rates for the approximation error of the cost can be obtained. Our algorithm significantly reduces the computational complexity of finding controls with asymptotically optimal cost. Numerical experiments using artificial neural networks as well as radial basis function networks illustrate the performance of our algorithm. Our approach can also be applied to stochastic control problems for high dimensional stochastic differential equations and more general stochastic partial differential equations., Comment: 12 figures

This is a book on weighted tree automata. We present the basic definitions and some of the important results in a coherent form with full proofs. The concept of weighted tree automata is part of Automata Theory and it touches the area of Universal Algebra. It originated from two sources: weighted string automata and finite-state tree automata.