13979 research outputs found
Sort by
A Novel Approach for Identification and Linking of Short Quotations in Scholarly Texts and Literary Works
Full text linkWe present two approaches for the identification and linking of short quotations between scholarly works and literary works: ProQuo, a specialized pipeline, and ProQuoLM, a more general language model based approach. Our evaluation shows that both approaches outperform a strong baseline and the overall performance is on the same level. We compare the performance of ProQuoLM on texts with and without (page) reference information and find that reference information is not used. Based on our findings, we propose the following steps for future improvements: further analysis of the influence of a bigger context window for better handling of long distance references and the introduction of positional information of the literary work so that reference information can be utilized by ProQuoLM
InvBERT: Reconstructing Text from Contextualized Word Embeddings by inverting the BERT pipeline
Full text linkDigital Humanities and Computational Literary Studies apply automated methods to enable studies on large corpora which are not feasible by manual inspection alone. However, due to copyright restrictions, the availability of relevant digitized literary works is limited. Derived Text Formats (DTFs) have been proposed as a solution. Here, textual materials are transformed in such a way that copyright-critical features are removed, but that the use of certain analytical methods remains possible. Word embeddings produced by transformer-encoders are promising candidates for DTFs because they allow for state-of-the-art performance on analytical tasks. However, in this paper we demonstrate that under certain conditions the reconstruction of the original text from token representations becomes feasible. Our attempts to invert BERT suggest, that publishing the encoder together with the contextualized embeddings is critical, since it allows to generate data to train a decoder with a reconstruction accuracy sufficient to violate copyright laws
Improvement of the stability with a TRT‐collision scheme in a Lattice Boltzmann method for elastic solids
Full text linkIn recent years, Lattice Boltzmann (LB) schemes have been adapted to the simulation of elastodynamics. The resulting methods show promising characteristics, especially the computational efficiency and the scaling to large systems. A constitutive model is needed for solids, which is captured by a source term in the LB formulation. However, this approach leads to problems with the numerical stability for a large regime of material parameters. In this publication, a linear elastic material model is regarded. This work regards the underlying reasons for instabilities. As a remedy, a two‐relaxation‐time collision operator is employed. Effects on the stability are examined through error measures of a dynamical simulation model
Drying of Soft Colloidal Films
Full text linkThin films made of deformable micro‐ and nano‐units, such as biological membranes, polymer interfaces, and particle‐laden liquid surfaces, exhibit a complex behavior during drying, with consequences for various applications like wound healing, coating technologies, and additive manufacturing. Studying the drying dynamics and structural changes of soft colloidal films thus holds the potential to yield valuable insights to achieve improvements for applications. In this study, interfacial monolayers of core‐shell (CS) microgels with varying degrees of softness are employed as model systems and to investigate their drying behavior on differently modified solid substrates (hydrophobic vs hydrophilic). By leveraging video microscopy, particle tracking, and thin film interference, this study shed light on the interplay between microgel adhesion to solid surfaces and the immersion capillary forces that arise in the thin liquid film. It is discovered that a dried replica of the interfacial microstructure can be more accurately achieved on a hydrophobic substrate relative to a hydrophilic one, particularly when employing softer colloids as opposed to harder counterparts. These observations are qualitatively supported by experiments with a thin film pressure balance which allows mimicking and controlling the drying process and by computer simulations with coarse‐grained models
Modeling Phase Separation in Grain‐Fluid Mixture Flows by a Depth‐Averaged Approach With Dilatancy Effects
Full text linkIn this work, we propose a comprehensive two‐layer depth‐averaged model to study the dynamic behavior of grain‐fluid mixtures, which considers the granular dilatancy effects and the different frictional rheologies of grains in different states. Unlike single‐phase flows, not only the interaction between granular and fluid phases significantly influence the dynamics of mixtures, but also the phase separation, so that different flow regimes can occur. These include five different possible regimes: two‐layer regimes of (a) under‐saturated mixture and (b) over‐saturated mixture as well as single‐layer regimes of (c) saturated mixture, (d) pure grains and (e) pure fluid. Most depth‐averaged models in previous studies have considered only one of these flow regimes. The present model is an improved and integrated version of these depth‐averaged models. Taking into account that the pure grains and pure fluid in the upper layer, which occur in the regimes of the under‐saturated and over‐saturated grain‐fluid mixtures, respectively, exhibit different flow features than in the lower layer of the saturated mixture, we use a two‐phase two‐layer depth‐averaged model to describe these regimes. This proposed model is possibly the first to employ a two‐layer structure to describe all possible different flow regimes simultaneously. The proposed model is then solved numerically using a high‐resolution central‐upwind scheme and shows its ability to handle different flow regimes. To demonstrate the robustness of the numerical implementation and to evaluate the performance of the model, the numerical results are compared with several experiments reported in the literature, showing a certain qualitative agreement
Behavior of a phase field model for wetting on structured surfaces
Full text linkTechnical surfaces are generally not perfectly smooth, but usually exhibit roughness. Additionally, as micro production techniques continue to evolve, geometrical surface structures at the micro scale can be manufactured, which presents a challenge for wetting models that are commonly formulated for smooth surfaces. A phase field model for surface wetting, proposed by Diewald et al., serves as a basis for this investigation. On smooth surfaces, the width of the gas‐liquid interface can be widened for scale bridging purposes, as this allows for accurate computations on coarse grids. However, it was observed that this interface scaling impacts the results when the model is applied to rough surfaces, and therefore a free choice of the interface width is not always sensible. In this work, the ability of the model to deal with sinusoidally shaped surfaces is investigated. An Allen–Cahn evolution equation is used to determine static equilibrium configurations for droplets on such surfaces
Influence of Grain Size on the Electrochemical Performance of Li₇₋₃ₓLa₃Zr₂AlₓO₁₂ Solid Electrolyte
Full text linkContemporary Li‐ion batteries are facing substantial challenges like safety and limited energy density. The development of all‐solid‐state battery cells mitigates safety hazards and allows the use of Li‐metal anodes increasing energy density. Garnet‐type solid electrolytes can be vital to achieving an all‐solid‐state cell and an understanding of the influence of its microstructure on the electrochemical performance is crucial for material and cell design. In this work the influence of grain size on the Li‐ion conductivity of Li₇₋₃ₓLa₃Zr₂AlₓO₁₂ (x=0.22) is presented. The synthesis and processing procedure allows changing the ceramic grain size, while maintaining the same synthesis parameters, eliminating influences of the synthesis on grain boundary composition. Field assisted sintering technology is a powerful method to obtain dense, fine‐grained ceramics with an optimal grain size of 2–3 μm, where the conductivity is double that of the counterpart (0.7 μm). A total Li‐ion conductivity of 0.43 mS cm⁻¹ and an activation energy of 0.36 eV were achieved. The oxide‐based all‐solid‐state battery cell combining the garnet‐type electrolyte, a Li‐metal anode and a thin‐film LiCoO₂ cathode was assembled and cycled at room temperature for 90 hours. This represents a proof of concept, for the application of oxide‐based electrolytes at ambient temperatures
A Data-Driven Framework for Car-in-the-Loop Test Bench Control and Optimization
Full text linkThe mega-trends in automotive industry, from advanced driver assistance systems to electrification and autonomous driving, are driving the further development of the "Road-to-Rig" concept. This results in the emergence of the multi-dynamic complete vehicle test benches, which simulate dynamic driving scenarios in different degrees of freedom simultaneously under laboratory conditions. The Car-in-the-Loop (CiL) concept, which
has its advantages in terms of compactness and modularity, belongs to this class of test benches.
For the purpose of proof-of-concept, a quarter-vehicle CiL prototype was built in the laboratory of the Institute for Mechatronic Systems at the Technical University of Darmstadt. Previous work on the CiL control demonstrates the potential of the model-based controllers for the control of the test bench longitudinal dynamics. However, the control strategy that is implemented results in undesired oscillations on the side shaft of the vehicle under test (VUT) during dynamic driving scenarios that capture the tire slip and load change phenomena. Moreover, the drive torque from the VUT, which is not necessarily available in practice, is assumed known.
To address these issues, a data-driven framework called AI4CiL (A Data-drIven framework for Car-in-the-Loop control and optimization) is proposed in this work. It combines the test bench control with machine learning techniques. In AI4CiL, different model-based controllers that have the potential to improve the reference tracking quality and disturbance rejection behaviour are integrated. A generalized performance evaluation function is suggested, serving as the objective function for the core algorithm. To be specific, in AI4CiL, a Gaussain Process provides a model that maps the optimization variables to the
performance evaluation function. Bayesian Optimization utilizes this model to acquire the next iteration location which is supposed to maximize the performance improvement.
In this work, a variety of AI4CiL applications are implemented for the CiL control and optimization. Among others, AI4CiL is used to learn performance-driven models, which are optimized directly under closed-loop conditions with the aim to improve the control performance. Furthermore, safeOpt-type algorithm is integrated into AI4CiL to
safely optimize the controllers directly on the CiL prototype. The optimized controllers are experimentally validated with dynamic driving scenarios on the CiL prototype. It is demonstrated that different controllers are best suitable for different scenarios. For example, the LQG controller shows the best disturbance rejection behaviour, without inducing oscillations on the side shaft of the VUT. Finally, all the controllers demonstrate good reproducibility and robustness by performing tests on the CiL prototype at different steering positions.
In summary, with the help of AI4CiL, the control performance of the CiL concept for dynamic driving scenarios is significantly improved. Further development of controllers or optimization algorithms can be easily integrated into the existing structure. In addition, due to its practicability and flexibility, the proposed framework AI4CiL can be transferred for the control and optimization of other multi-dynamic complete vehicle test benches
Code of Courage: A Comic About Digital Security for Activists
Full text linkSecurity is paramount - online and offline.
Maria, Alex, Sarah, and Daniel are four friends deeply committed to human rights activism in a country grappling with increasing authoritarianism. Their efforts to organize and amplify their message on social media are met with relentless challenges: internet shutdowns, hacked accounts, direct threats, and constant surveillance - tactics aimed at silencing dissent and instilling fear. Daniel's arrest for this involvement in the protests serves as a stark reminder of the dangers they all face.
Follow Maria as she strives to learn more about digital safety and what is means for activists in four short comic stories
Proximity-based screening of a class A/B hybrid G protein-coupled receptor
Full text linkAmong all human cell surface receptors G protein-coupled receptors (GPCRs) represent the largest group. The rhodopsin-like GPCRs, or class A GPCRs, represent the majority of all GPCRs and are participating in a variety of different signalling pathways in the human body. Because of their important functions and their role in various diseases class A GPCRs serve as crucial targets for drug development. The high demand for ligands for GPCRs is confronted with the problem of the detection limit for weak ligands. This makes it more difficult to find potentially interesting lead structures for GPCRs. With the biomimetic screening approach by Devigny et al. this problem was addressed for class B GPCRs. In this screening approach, the suggested two-site-/two-step activation mechanism of class B GPCRs was utilized to screen a peptide conjugate library for the corticotropin releasing factor receptor 1 (CRHR₁). Thereby initial weak binders for CRHR₁ could be detected and further optimized. In this work a system of a class A/B hybrid GPCR was designed consisting of the model class A GPCR endothelin receptor B (ET_B) and the N-terminal extracellular domain of the CRHR₁. Peptide probe sequences conjugated to a CRHR₁-extracellular domain binding ligand, termed ‘carrier peptide’, were screened for activation of the hybrid receptors, making use of the proximity effect caused by the carrier peptide. The new approach resulted in a 20-fold activity improvement. This allowed the detection of weak peptide probes that were inactive without conjugation to the CRHR₁-extracellular domain binding ligand. In addition, a structure-activity relationship (SAR) study for the C-terminal part of Endothelin-1 (ET-1) that is crucial for activation of ET_B was performed. Towards this an ET-1 derived peptide library was synthesized, conjugated to the CRHR₁ extracellular domain-binding ligand and tested towards the ability to activate the CRHR₁-ET_B hybrid receptor. Validation of generated hits offered more detailed insights in receptor-ligand interactions necessary for receptor activation. Finally, a series of class A/B hybrid GPCRs was developed and tested for activity with orthosteric agonists. Despite the intensive study of the GPCR protein family in the past, ongoing GPCR drug research persists due to the lack of selective ligands and the need for deorphanization of receptors. Tools that facilitate the screening for class A GPCR ligands, enabling the detection of weak activating fragments as described in this work, are important to address these open questions