727 research outputs found
Sort by
Encoder-Decoder Model for Semantic Role Labeling
No full textAbstract (Daza & Frank 2019):
We propose a Cross-lingual Encoder-Decoder model that simultaneously translates and generates sentences with Semantic Role Labeling annotations in a resource-poor target language. Unlike annotation projection techniques, our model does not need parallel data during inference time. Our approach can be applied in monolingual, multilingual and cross-lingual settings and is able to produce dependency-based and span-based SRL annotations. We benchmark the labeling performance of our model in different monolingual and multilingual settings using well-known SRL datasets. We then train our model in a cross-lingual setting to generate new SRL labeled data. Finally, we measure the effectiveness of our method by using the generated data to augment the training basis for resource-poor languages and perform manual evaluation to show that it produces high-quality sentences and assigns accurate semantic role annotations. Our proposed architecture offers a flexible method for leveraging SRL data in multiple languages.</p
Sensor-Based Measurements in Paraplegia: Classified References from a Systematic Review
No full textThis dataset contains the results (publication references) of the systematic review "Current Use of Sensor-Based Measurements for Paraplegics", presented at MIE 2020, Geneva
Converter for content-to-head style syntactic dependencies
No full textA set of Python scripts that convert function-head style encodings in dependency treebanks in a content-head style encoding (as used in the UD treebanks) and vice versa (for adpositions, copula and coordination). For more information, see (Rehbein, Steen, Do & Frank 2017)
German causal language annotations and lexicon (verbs, nouns, prepositions) (DE)
No full textAnnotations of causal verbs, nouns and prepositions in context and lexicon file for causal verbs, nouns and prepositions
Cataloging Cultural Objects (CCO) – The CCO Commons examples in VRA Core 4 XML
No full text“Cataloging Cultural Objects - a Guide to Describing Cultural Works and Their Images” (CCO) provides a data content standard for catalogers of cultural heritage. It is a guidebook for how to populate data elements and where to apply controlled vocabulary standards. The guide is further supported by a larger number of examples in text format provided on the CCO Commons webpage. A project of the Heidelberg Research Architecture focused on encoding all CCO Commons examples in machine readable format using the "strict" version of the VRA Core 4 XML metadata schema. The records make use of the multilingual extensions of the VRA Core 4, developed by the HRA. We provide the full XML data set as example for the implementation of VRA Core 4 XML and a use case for its adaptation to multi-lingual and non-Latin script metadata
Supplementary information to 'Defining pottery use and exploitation of natural products at Clairvaux XIV during the Middle Neolithic'
No full textInvestigating pottery function requires optimal preservation of both ceramic vessels and organic residues related to vessel use. If these criteria are met then opportunities emerge for combining macroscopic observations, typological classification, and capacity measurements and molecular and isotopic proxies (GC, GC-MS, and GC-C-IRMS) to explore vessel use and resource exploitation. The exceptional ceramic assemblage of hundreds of vessels excavated from the Middle Neolithic site of Clairvaux XIV (Jura, France, IVth millennium B.C.) offered such an opportunity, with a wide diversity of vessel shapes and extraordinary preservation of lipids. Interrogating the lipid residues together with typological analyses revealed a complex system of ceramic vessel use. Despite the scarcity of aquatic products, a wide diversity of natural substances has been detected: ruminant adipose fats, dairy products, plant substances, and beeswax. These commodities were processed in specific vessels with three main different functions:
cooking vessels of various volumes, serving dishes for animal products, and dedicated vessels for dairy product processing and consumption
A Test of Information Aversion
No full textThe standard Bayesian model implies that information can never have a negative value. We put this implication to the proof. Our paper provides the first test of the value (positive or negative) of information under uncertainty. We show that the ``Bayesian implication'' stands in conflict with the information-averse behavior that is revealed in our experiment. This behavior demonstrates that the value of truthful and unambiguous information may indeed be negative. Our findings complement predictions from recent theoretical work in showing that negative value of information correlates with ambiguity aversion. This highlights the importance of counseling for decision-making under uncertainty
Dissecting intratumor heterogeneity of nodal B cell lymphomas at transcriptional, genetic, and drug response levels
No full textTumour heterogeneity encompasses both the malignant cells and their microenvironment. While heterogeneity between individual patients is known to affect the efficacy of cancer therapy, most personalized treatment approaches do not account for intratumor heterogeneity. We addressed this issue by studying the heterogeneity of nodal B cell lymphoma by single cell RNA-sequencing and transcriptome-informed flow cytometry. We identified transcriptionally distinct malignant subpopulations and compared their drug response and genomic profiles. Malignant subpopulations of the same patient responded strikingly different to anti-cancer drugs ex vivo, which recapitulated subpopulation-specific drug sensitivity during in vivo treatment. Infiltrating T cells represented the majority of non-malignant cells, whose gene expression signatures were similar across all donors, whereas the frequencies of T cell subsets varied significantly between the donors. Our data provide insights into the heterogeneity of nodal B cell lymphoma and highlight the relevance of intratumor heterogeneity for personalized cancer therapy
Direct control of high magnetic fields for cold atom experiments based on NV centers [Dataset]
No full textIn atomic physics experiments, magnetic fields allow to control the interactions between atoms, eg. near Feshbach resonances, or by employing spin changing collisions. The magnetic field control is typically performed indirectly, by stabilizing the current of Helmholtz coils producing the large bias field. Here, we overcome the limitations of such an indirect control through a direct feedback scheme, which is based on nitrogen-vacancy centers acting as a sensor. This allows us to measure and stabilize magnetic fields of 46.6 G down to 1.2 mG RMS noise, with the potential of reaching much higher field strengths. Because the magnetic field is measured directly, we reach minimum shot-to-shot fluctuations of 0.32(4) ppm on a 22 minute time interval, ensuring high reproducibility of experiments. This approach extends the direct magnetic field control to high magnetic fields, which could enable new precise quantum simulations in this regime
