Fractal Systems (Canada)

Visual prompting aims to enhance the performance of vision-language models (VLMs), which, despite their remarkable capabilities, often struggle with dense, detailed images, leading to incorrect answers or hallucinations. We propose Visual Prompting Through Image Mines , a novel algorithm that leverages attention patterns from a base VLM to generate image crops for improved visual grounding. Specifically, we extract attention values from output text tokens in the LLaVA-8B model, overlay them onto image patches to create an attention graph, and apply a modified breadth-first search (BFS) to identify key regions as image crops. Using the SigLIP model, we refine these regions into Image Mines, retaining only the most relevant crops. Our approach supports both single-image and multi-image inference setups, delivering superior performance. It consistently outperforms existing visual prompting methods in nearly all scenarios.

In the current climate of sustainable animal agriculture, nutritional strategies that support fattening swine growth performance and bone mineralization whilst reducing environmental impacts are much sought after. This study evaluated the effect of supplementing 25(OH)D3 with triterpenoids to a Ca-reduced diet containing phytase during the grower-finisher phase. Growth performance, bone composition, plasma metabolites and muscle gene expression were evaluated. Sixty crossbreed boar pigs (initial body weight (BW) 42.0 ± 5.1 kg at 12 wk of age) were assigned to three treatments with 20 pigs/treatment in a completely randomized design. Treatments comprised: 1) a standard commercial grower-finisher diet (positive control (PC)) containing 1,500 IU/kg vitamin D3 [3,585 kcal/kg digestible energy, 16.19% CP, 0.70% Ca, 0.29% standardized total tract digestible P]; 2) a negative control (NC) based on the PC with reduction in Ca and P (minus 30% and 10%, respectively); 3) the NC with vitamin D3 replaced by a commercially available compounds combination containing 25(OH)D3 and triterpenoids, dosed at 500 mg per kg of feed (TRT). All diets were provided ad libitum for 7 wk, and feed intake was recorded individually via electronic feeder stations. For the overall period, average daily gain and average daily feed intake were increased (P < 0.05) in TRT vs. NC or PC (+ 13.0% and + 8.3%, respectively, vs. NC); final BW was 7.8% higher vs. NC (+ 5.2% vs. PC; P < 0.05). Whole-body DXA-scanning at 19 wk of age showed that bone mineral density, content and percentage were reduced in NC vs. PC and equivalent to PC in TRT. Plasma 25(OH)D3 and P levels were raised in TRT (+ 33 ng/ml or 2.6-fold and + 0.55 mg/dL or 5.9%, respectively, vs. NC). The combination of 25(OH)D3 with triterpenoids was found to activate several biological pathways involved in muscle growth, including pathways that activate mTOR, a key central regulator of cell metabolism, growth, proliferation and survival when the gene expression was measured in the muscle tissue at 19 wk of age. These results suggest that the dietary combination of 25(OH)D3 with triterpenoids has the potential for use, alongside phytase, in supporting a reduction in Ca and P in the diet to reduce nutrient waste and improve the sustainability of production by promoting muscle growth and maintaining bone composition.

Nondestructive testing (NDT) of composite materials is of paramount importance to the aerospace industry. Several NDT methods have been adopted for the inspection of components during production and all through the aircraft service life, with infrared thermography (IRT) techniques, such as line scan thermography (LST) and pulsed thermography (PT), gaining popularity thanks to their rapidity and versatility. On one hand, LST is an attractive solution for the fast inspection of large and complex geometry composite parts during production. On the other hand, PT can be employed for the characterization of composite materials, e.g., the determination of thermal diffusivity and defect depth estimation. In this study, the use of LST with an uncooled microbolometer camera is explored for the identification of artificially produced porosity and barely visible impact damage (BVID) on academic samples. The performance of LST is quantitatively assessed with respect to PT (considered the gold standard in this case) using a high-definition cooled camera through the contrast-to-noise ratio (CNR) criterium. It is concluded that, although in most cases the measured CNR values were higher for PT than for LST (as expected since a high-definition camera and longer acquisition times were used), the majority of the defects were clearly detected (CNR ≥ 2.5) by LST without the need of advanced signal processing, proving the suitability of LST for the inspection of aerospace composite components. Furthermore, the deepest defect investigated herein (z ≈ 3 mm) was detected solely by LST combined with signal processing and spatial filtering (CNR = 3.6) and not by PT (since pulse heating was not long enough for this depth). In addition, PT was used for the determination of the thermal diffusivity of all samples and the subsequent depth estimation of porosity and damaged areas by pulsed phase thermography (PPT).

Abstract Contact tracing forms a crucial part of the public-health toolbox in mitigating and understanding emergent pathogens and nascent disease outbreaks. Contact tracing in the United States was conducted during the pre-Omicron phase of the ongoing COVID-19 pandemic. This tracing relied on voluntary reporting and responses, often using rapid antigen tests (with a high false negative rate) due to lack of accessibility to PCR tests. These limitations, combined with SARS-CoV-2’s propensity for asymptomatic transmission, raise the question “how reliable was contact tracing for COVID-19 in the United States”? We answered this question using a Markov model to examine the efficiency with which transmission could be detected based on the design and response rates of contact tracing studies in the United States. Our results suggest that contact tracing protocols in the U.S. are unlikely to have identified more than 1.65% (95% uncertainty interval: 1.62%-1.68%) of transmission events with PCR testing and 0.88% (95% uncertainty interval 0.86%-0.89%) with rapid antigen testing. When considering an optimal scenario, based on compliance rates in East Asia with PCR testing, this increases to 62.7% (95% uncertainty interval: 62.6%-62.8%). These findings highlight the limitations in interpretability for studies of SARS-CoV-2 disease spread based on U.S. contact tracing and underscore the vulnerability of the population to future disease outbreaks, for SARS-CoV-2 and other pathogens.

The duration of natural immunity in response to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a matter of some debate in the literature at present. For example, in a recent publication characterizing SARS-CoV-2 immunity over time, the authors fit pooled longitudinal data, using fitted slopes to infer the duration of SARS-CoV-2 immunity. In fact, such approaches can lead to misleading conclusions as a result of statistical model-fitting artifacts. To exemplify this phenomenon, we reanalyzed one of the markers (pseudovirus neutralizing titer) in the publication, using mixed-effects modeling, a methodology better suited to longitudinal datasets like these. Our findings showed that the half-life was both longer and more variable than reported by the authors. The example selected by us here illustrates the utility of mixed-effects modeling in provide more accurate estimates of the duration and heterogeneity of half-lives of molecular and cellular biomarkers of SARS-CoV-2 immunity.

Earth and Space Science Open Archive posterOpen AccessYou are viewing the latest version by default [v1]SKB Task Force GWFTS: Task 9 - Increasing the realism of solute transport modelling in fractured mediaAuthorsBjörnGyllingiDPaoloTrincheroiDJosepSoleriDJamesCrawfordKerstiNilssoniDBillLanyonJan-OlofSelroosiDAnttiPoteriSee all authors Björn GyllingiDCorresponding Author• Submitting AuthorGylling GeoSolutionsiDhttps://orcid.org/0000-0002-2464-6725view email addressThe email was not providedcopy email addressPaolo TrincheroiDAmphos21iDhttps://orcid.org/0000-0003-1351-2788view email addressThe email was not providedcopy email addressJosep SoleriDIDAEA-CSICiDhttps://orcid.org/0000-0003-0741-249Xview email addressThe email was not providedcopy email addressJames CrawfordKemakta Konsult ABview email addressThe email was not providedcopy email addressKersti NilssoniDGeosigma ABiDhttps://orcid.org/0000-0001-7140-5863view email addressThe email was not providedcopy email addressBill LanyonFracture Systems Ltd.view email addressThe email was not providedcopy email addressJan-Olof SelroosiDSKB (Svensk Kärnbränslehantering AB)iDhttps://orcid.org/0000-0001-9251-5367view email addressThe email was not providedcopy email addressAntti PoteriPosiva Oyview email addressThe email was not providedcopy email address

Thousands of videos are posted online every day. The affordability of video editing tools and social networks has facilitated the creation and spread of videos carrying disinformation, i.e. fake videos. Previous attempts to categorize disinformation have focused on content analysis and ascertaining the intention of creators. To extend these approaches, it is beneficial to incorporate the perspective of other fields that study the trustworthiness of records, such as archival science, to help detect and categorize fake videos. This paper proposes to leverage archival science in combination with computer engineering to devise a new framework for detecting and categorizing fake videos. In doing so, the paper offers a case study of the way in which Computational Archival Science, which blends archival and computational thinking, can be used to contribute to a novel approach towards solving the problem of fake videos.

.

Urucum (urucum ou Bixa OrellanaL.), uma árvore pequena da família Bixaceae, é nativa da América do Sul e Central e é difundida nos trópicos. Embora a bixina pigmento natural da semente do urucum, representando 80% de todos carotenoides presentes, possui cadeia mais curta (25 C) do que os habituais carotenoides (C 40), a bixina também possui uma cadeia de duplas ligações conjugadas, o que lhe confere algumas características especiais. Das metodologias utilizadas para extração, o Soxhlet é um dos que permite um rendimento maior. Esta extração resulta em produtos distintos quanto à qualidade e concentração de ativos.As sementes de duas variedades de Bixa orellana cultivadas em Morretes Paraná analisadas mostraram que a variedade vermelha apresentaram concentrações após a extração, fração hexano 6,59%, clorofórmio 5,07% de bixina isolada, acetato de eitla 0,75% e fração hidroalcólica 5,75%, enquanto a variedade verde fração hexano 6,79%, clorofórmio 3,28% de bixina isolada, acetato de eitla 0,87% e fração hidroalcólica 4,8%. Observou-se a presença cristais 1,378%, que foram identificados com sendo sacarose na variedade vermelha após algum tempo em repouso.