Search for: All records

Compositional models represent patterns with hierarchies of meaningful parts and subparts. Their ability to characterize high-order relationships among body parts helps resolve low-level ambiguities in human pose estimation (HPE). However, prior compositional models make unrealistic assumptions on subpart-part relationships, making them incapable to characterize complex compositional patterns. Moreover, state spaces of their higher-level parts can be exponentially large, complicating both inference and learning. To address these issues, this paper introduces a novel framework, termed as Deeply Learned Compositional Model (DLCM), for HPE. It exploits deep neural networks to learn the compositionality of human bodies. This results in a novel network with a hierarchical compositional architecture and bottom-up/top-down inference stages. In addition, we propose a novel bone-based part representation. It not only compactly encodes orientations, scales and shapes of parts, but also avoids their potentially large state spaces. With significantly lower complexities, our approach outperforms state-of-the-art methods on three benchmark datasets. 

Nanocellulose is increasingly considered for applications; however, the fibrillar nature, crystalline phase, and surface reactivity of these high aspect ratio nanomaterials need to be considered for safe biomedical use. Here a comprehensive analysis of the impact of cellulose nanofibrils (CNF) and nanocrystals (CNC) is performed using materials provided by the Nanomaterial Health Implications Research Consortium of the National Institute of Environmental Health Sciences. An intermediary length of nanocrystals is also derived by acid hydrolysis. While all CNFs and CNCs are devoid of cytotoxicity, 210 and 280 nm fluorescein isothiocyanate (FITC)‐labeled CNCs show higher cellular uptake than longer and shorter CNCs or CNFs. Moreover, CNCs in the 200–300 nm length scale are more likely to induce lysosomal damage, NLRP3 inflammasome activation, and IL‐1β production than CNFs. The pro‐inflammatory effects of CNCs are correlated with higher crystallinity index, surface hydroxyl density, and reactive oxygen species generation. In addition, CNFs and CNCs can induce maturation of bone marrow–derived dendritic cells and CNCs (and to a lesser extent CNFs) are found to exert adjuvant effects in ovalbumin (OVA)‐injected mice, particularly for 210 and 280 nm CNCs. All considered, the data demonstrate the importance of length scale, crystallinity, and surface reactivity in shaping the innate immune response to nanocellulose.

 

Carbon nanotubes (CNTs) exhibit a number of physicochemical properties that contribute to adverse biological outcomes. However, it is difficult to define the independent contribution of individual properties without purified materials. A library of highly purified single‐walled carbon nanotubes (SWCNTs) of different lengths is prepared from the same base material by density gradient ultracentrifugation, designated as short (318 nm), medium (789 nm), and long (1215 nm) SWCNTs. In vitro screening shows length‐dependent interleukin‐1β (IL‐1β) production, in order of long > medium > short. However, there are no differences in transforming growth factor‐β1 production in BEAS‐2B cells. Oropharyngeal aspiration shows that all the SWCNTs induce profibrogenic effects in mouse lung at 21 d postexposure, but there are no differences between tube lengths. In contrast, these SWCNTs demonstrate length‐dependent antibacterial effects onEscherichia coli, with the long SWCNT exerting stronger effects than the medium or short tubes. These effects are reduced by Pluronic F108 coating or supplementing with glucose. The data show length‐dependent effects on proinflammatory response in macrophage cell line and antibacterial effects, but not on collagen deposition in the lung. These data demonstrate that over the length scale tested, the biological response to highly purified SWCNTs is dependent on the complexity of the nano/bio interface.

 

Abstract Many measurements at the LHC require efficient identification of heavy-flavour jets, i.e. jets originating from bottom (b) or charm (c) quarks. An overview of the algorithms used to identify c jets is described and a novel method to calibrate them is presented. This new method adjusts the entire distributions of the outputs obtained when the algorithms are applied to jets of different flavours. It is based on an iterative approach exploiting three distinct control regions that are enriched with either b jets, c jets, or light-flavour and gluon jets. Results are presented in the form of correction factors evaluated using proton-proton collision data with an integrated luminosity of 41.5 fb -1 at  √s = 13 TeV, collected by the CMS experiment in 2017. The closure of the method is tested by applying the measured correction factors on simulated data sets and checking the agreement between the adjusted simulation and collision data. Furthermore, a validation is performed by testing the method on pseudodata, which emulate various mismodelling conditions. The calibrated results enable the use of the full distributions of heavy-flavour identification algorithm outputs, e.g. as inputs to machine-learning models. Thus, they are expected to increase the sensitivity of future physics analyses.