2972 research outputs found
Sort by
Bayesian Hierarchical Models can Infer Interpretable Predictions of Leaf Area Index From Heterogeneous Datasets
Get PDFEnvironmental scientists often face the challenge of predicting a complex phenomenon from a heterogeneous collection of datasets that exhibit systematic differences. Accounting for these differences usually requires including additional parameters in the predictive models, which increases the probability of overfitting, particularly on small datasets. We investigate how Bayesian hierarchical models can help mitigate this problem by allowing the practitioner to incorporate information about the structure of the dataset explicitly. To this end, we look at a typical application in remote sensing: the estimation of leaf area index of white winter wheat, an important indicator for agronomical modeling, using measurements of reflectance spectra collected at different locations and growth stages. Since the insights gained from such a model could be used to inform policy or business decisions, the interpretability of the model is a primary concern. We, therefore, focus on models that capture the association between leaf area index and the spectral reflectance at various wavelengths by spline-based kernel functions, which can be visually inspected and analyzed. We compare models with three different levels of hierarchy: a non-hierarchical baseline model, a model with hierarchical bias parameter, and a model in which bias and kernel parameters are hierarchically structured. We analyze them using Markov Chain Monte Carlo sampling diagnostics and an intervention-based measure of feature importance. The improved robustness and interpretability of this approach show that Bayesian hierarchical models are a versatile tool for the prediction of leaf area index, particularly in scenarios where the available data sources are heterogeneous
Quality-Aware Compressed Sensing for Distributed Precision Agriculture Systems
No full textAround 2006 the signal processing community was thrilled when the concept of Compressed Sensing was brought forward. While according to the long-established Nyquist-Shannon theorem, the sampling rate for signals must be at least twice as high as the highest relevant frequency in the signal, with Compressed Sensing lower average sampling rates become suitable. As a more general toolkit, Compressed Sensing allows for merging sensing and compression in a single step. In other words, only a tiny amount of data needs to be sensed but a huge amount of information can be reconstructed from this data. While Compressed Sensing already has been successfully applied in some areas, mainly in medical scanning where it helps to reduce the exposure to radiation, adaptation to new application areas is relatively slow. We identify two causes that slow down the adaption and contribute to overcoming these obstacles: firstly, the adaption of Compressed Sensing requires interdisciplinary thinking even more than for many other new technologies: Compressed Sensing itself comes from the field of signal processing, the application adds a second field. The automatized processing of the data always touches the field of computer science and lastly, designing Compressed Sensing solutions often requires a modification or redesign of measurement hardware, touching the fields of electronics and sometimes mechanics. We address this issue by supplying a more structured approach for designing Compressed Sensing solutions. Secondly, Compressed Sensing usually performs a lossy compression on real world data. Not knowing the quality of the solution limits its usability.
We address this issue by supplying a list of potential metrics for assessing the quality of the solution and evaluate their performance for various datasets.
Along the way, we develop two Compressed Sensing solutions in the application area of Precision Agriculture: The first is Compressive Field Estimate (CFE), a method for improving the remote estimate of a scalar field based on limited data supplied by a moving probe such as a combine harvester.
The second is Multi- to Hyperspectral Sensor Network (M2HSN), a wireless sensor network that records light spectra at mediocre spectral resolution and allows for increasing the spectral resolution.
For the M2HSN, we discuss different designs of the sensor nodes and different approaches for increasing the resolution.
Those are simulatively evaluated on different datasets and in a real-world prototype
Biosecurity of Synthetic Viruses
Get PDFSynthetic methods in life sciences made rapid progress in the past few years, such as synthetic biology, synthetic genomics and synthetic virology. Synthetic viruses are genetically engineered viruses that are created to get deeper insights into the function of viruses, but they could also be used to develop live vaccines. Synthetic viruses are a small, but increasingly relevant area in virology. The open access to DNA sequence databases of many relevant viruses including smallpox and horsepox and the possibility to order DNA sequences commercially are the main biosecurity issues of synthetic virology. The open databases, the DNA synthesis firms and their clients are in the center of the biosecurity discussion and regulations. Extinct viruses could already be recreated as synthetic viruses solely on database information and commercially ordered synthetic DNA. The smallpox virus with its very high transmissibility and mortality is extinct and meanwhile new antivirals and vaccines are available which demonstrated their effect during the current monkeypox outbreak. However, the extinct horsepox virus that is closely related to smallpox could be re-created in 2017, i.e., it is technically possible to synthesize poxviruses and it is unlikely that the stockpiled antivirals and vaccines could control a new smallpox pandemic. Technical and financial hurdles for synthetic viruses remain high, but are already lower than in the past. Currently, tacit expert knowledge is still an important factor. Governments and scientific institutions need to detect earliest signs of critical developments. The paper gives a brief overview on synthetic biology, synthetic genomics and virology, the development of synthetic viruses, smallpox-related matters and the biosecurity measures
Intonation Patterns Used in Non-Neutral Statements by Czech Learners of Italian and Spanish: A Cross-Linguistic Comparison
No full textThe objective of the study is to contribute to our understanding of the acquisition of second language intonation by comparing L2 Italian and L2 Spanish as produced by L1 Czech learners. Framed within the L2 Intonation Learning theory, the study sheds light on which tonal events tend to be successfully learnt and why. The study examines different types of non-neutral statements (narrow focus, statements of the obvious, what-exclamatives), obtained by means of a Discourse Completion Task. The findings show that the two groups diverge significantly in producing the nuclear pitch accents L+H* (L2 Spanish) and (L+)H*+L (L2 Italian), which is indicative of a target-like realization in each language. However, the learners struggle with the acquisition of the target boundary tones HL% and L!H% in L2 Spanish and prenuclear pitch accents in both Romance varieties. It is speculated that this is due not only to difficulties in acquiring semantic or systemic dimensions, but also to perceptual salience and frequency effects. In addition, the study explores individual differences and reveals no significant effects of the time spent in an L2-speaking country, the age of learning and the amount of active use of a foreign language on accuracy in L2 production
A continuous time meta-analysis of the relationship between conspiracy beliefs and individual preventive behavior during the COVID-19 pandemic
Get PDFIn several longitudinal studies, reduced willingness to show COVID-19-related preventive behavior (e.g., wearing masks, social distancing) has been partially attributed to misinformation and conspiracy beliefs. However, there is considerable uncertainty with respect to the strength of the relationship and whether the negative relationship exists in both directions (reciprocal effects). One explanation of the heterogeneity pertains to the fact that the time interval between consecutive measurement occasions varies (e.g., 1 month, 3 months) both between and within studies. Therefore, a continuous time meta-analysis based on longitudinal studies was conducted. This approach enables one to examine how the strength of the relationship between conspiracy beliefs and COVID-19 preventive behavior depends on the time interval. In total, 1035 correlations were coded for 17 samples (N = 16,350). The results for both the full set of studies and a subset consisting of 13 studies corroborated the existence of reciprocal effects. Furthermore, there was some evidence of publication bias. The largest cross-lagged effects were observed between 3 and 6 months, which can inform decision-makers and researchers when carrying out interventions or designing studies examining the consequences of new conspiracy theories
Progression in cognitive-affective research by increasing ecological validity: A series of Virtual Reality studies.
No full textThe ultimate aim of psychological research is to disentangle everyday human functioning. Achieving this goal has always been limited by the necessity of balancing experimental control and ecological validity. Recent technical advances, however, reduce this trade-off immensely, perhaps even rendering it void: Sophisticated virtual reality (VR) systems provide not only high experimental control but also multidimensional and realistic stimuli, tasks, and experimental setups. Yet prior to applying VR as a standalone experimental method, an empirical foundation for its application needs to be established.
To this end, this dissertation aims to shed light on whether and which changes in cognitive-affective standard findings result from increasing the ecological validity by means of VR paradigms. The four empirical studies included in this dissertation focus either on the affective or mnemonic processes and mechanisms occurring under immersive VR conditions compared to conventional laboratory setups. Study 1.1 investigated whether the electrophysiological correlates of the approach/avoidance dimension differ depending on the mode of presentation, i.e., immersive VR footage or a virtual 2D desktop. Study 2 was extended by a behavioral component. Full-body responses were enabled within this paradigm to examine holistic fear responses and to put to the test whether the respective electrophysiological responses translate from keystrokes to natural responses. With respect to the retrieval of such immersive experiences, Study 1.2 aimed to replicate the memory superiority effect found for VR conditions compared to conventional conditions. The generalizability of this effect will be examined using complex, multimodal scenes. Going one step further, Study 3 differentiated the retrieval mechanisms underlying VR-based or conventional laboratory engrams on the electrophysiological level. The well-established theta old/new effect served as a benchmark to check whether cognitive processes obtained under conventional conditions translate to VR conditions.
The results of these studies are discussed with respect to whether and how increasing ecological validity alters the standard findings expected on the basis of the previous research background. Special attention will be paid to the differences between conventional laboratory setups and sophisticated VR setups with the aim to identify possible sources of the obtained deviations from standard findings. Such changes in the findings that overlap and exceed all studies beyond their primary focus, whether emotional or mnemonic, are discussed in terms of embodied simulations and the predictive coding hypothesis. A shared mental 3D default space is proposed as a possible source of fundamental differences between conventional and VR-based research outcomes. In particular, it will be demonstrated that conventional research approaches and findings may not only be amplified but fundamentally altered when translated to VR paradigms
Sun protection and occupation: Current developments and perspectives for prevention of occupational skin cancer
No full textA substantial proportion of all reported occupational illnesses are constituted by skin cancer, making this disease a serious public health issue. Solar ultra-violet radiation (UVR) exposure is the most significant external factor in the development of skin cancer, for which the broad occupational category of outdoor workers has already been identified as high-risk group. Sun protection by deploying adequate technical, organizational, and person-related measures has to be understood as a functional aspect of workplace safety. To prevent skin cancers brought on by—typically cumulative—solar UVR exposure, outdoor workers must considerably lower their occupationally acquired solar UVR doses. Estimating cumulative sun exposure in outdoor workers requires consideration of the level of solar UVR exposure, the tasks to be done in the sun, and the employees' solar UVR preventive measures. Recent studies have highlighted the necessity for measures to enhance outdoor workers' sun protection behavior. In the coming decades, occupational dermatology is expected to pay increasing attention to sun protection at work. Also, the field of dermato-oncology will likely be concerned with sky-rocketing incidences of occupational skin cancers. The complete range of available alternatives should be utilized in terms of preventive actions, which seems pivotal to handle the present and future challenges in a purposeful manner. This will almost definitely only be possible if politicians' support is effectively combined with communal and individual preventive actions in order to spur long-term transformation
Classical versus Quantum Dynamics in Interacting Spin Systems
Get PDFThis dissertation deals with the dynamics of interacting quantum and classical spin models
and the question of whether and to which degree the dynamics of these models agree with
each other.
For this purpose, XXZ models are studied on different lattice geometries of finite size,
ranging from one-dimensional chains and quasi-one-dimensional ladders to two-dimensional
square lattices. Particular attention is paid to the high-temperature analysis of the temporal
behavior of autocorrelation functions for both the local density of magnetization (spin)
and energy, which are closely related to transport properties of the considered models. Due
to the conservation of total energy and total magnetization, the dynamics of such densities
are expected to exhibit hydrodynamic behavior for long times, which manifests itself in
a power-law tail of the autocorrelation function in time. From a quantum mechanical
point of view, the calculation of these autocorrelation functions requires solving the linear
Schrödinger equation, while classically Hamilton’s equations of motion need to be solved.
An efficient numerical pure-state approach based on the concept of typicality enables
circumventing the costly numerical method of exact diagonalization and to treat quantum
autocorrelation functions with up to N = 36 lattice sites in total.
While, in full generality, a quantitative agreement between quantum and classical dy-
namics can not be expected, contrarily, based on large-scale numerical results, it is
demonstrated that the dynamics of the quantum S = 1/2 and classical spins coincide, not
only qualitatively, but even quantitatively, to a remarkably high level of accuracy for all
considered lattice geometries. The agreement particularly is found to be best in the case
of nonintegrable quantum models (quasi-one-dimensional and two-dimensional lattice),
but still satisfactory in the case of integrable chains, at least if transport properties are
not dominated by the extensive number of conservation laws.
Additionally, in the context of disordered spin chains, such an agreement of the dynamics
is found to hold even in the presence of small values of disorder, while at strong disorder
the agreement is pronounced most for larger spin quantum numbers.
Finally, it is shown that a putative many-body localization transition within the one-
dimensional spin chain is shifted to stronger values of disorder with increasing spin
quantum number. It is concluded that classical or semiclassical simulations might provide
a meaningful strategy to investigate the quantum dynamics of strongly interacting quantum
spin models, even if the spin quantum number is small and far from the classical limit
Motivierungspotential von Unterrichtsmaterial. Anschluss-, Leistungs- und Machtthemen in Schulaufgaben und ihr Einfluss auf die Motivation und Leistung Lernender.
No full textTheoretischer Hintergrund. Motivation im schulischen Kontext ist ein wichtiger Prädiktor für das Lernverhalten und die Leistung. Häufig ist jedoch im Verlauf der Schulzeit ein Rückgang der schulischen Lernmotivation feststellbar. Unterrichtsmaterialien, wie Schulaufgaben, können genutzt werden, um die Motivation von Schüler:innen zu steigern, was sich wiederum positiv auf das Lernverhalten und die Leistung auswirken kann. Lernmaterial sollte Lernende besonders dann motivieren, wenn es persönlich relevante Themen beinhaltet, mit denen sie sich identifizieren können. Forschungsbefunden zufolge sind anschlussbezogene Themen, sowie leistungs- und macht- bzw. statusbezogene Themen für Schüler:innen persönlich relevant. Inhaltsanalysen von Schul- und Übungsbüchern (Deutsch und Mathematik) zeigen, dass die Anschluss-, Leistungs- und Machtthemen (Motivthemen) häufig zur Aufgabengestaltung verwendet werden. Gleichzeitig liefern diese Untersuchungen Hinweise auf eine alters- und geschlechtsstereotype Aufgabenformulierung in der Verwendung der genannten Themen. Bisher ist ungeklärt, ob Schulaufgaben mit motivthematischen Inhalten motivierender sind als Aufgaben ohne diese Inhalte und inwiefern Geschlecht und Alter der Schüler:innen dabei von Bedeutung sind. Überdies ist offen, ob sich eine motivthematische Aufgabengestaltung positiv auf die Leistung Lernender auswirkt. Im Rahmen der vorliegenden Dissertation wurde der Einfluss motivthematisch formulierter Schulaufgaben auf die Motivation und Leistung Lernender untersucht. Der erste und zweite Artikel überprüfte mittels experimenteller Feldstudien die potentiell motivierende Funktion von Anschluss-, Leistungs-, und Machtthemen in Schulaufgaben. Als Indikator für die Motivation wurde dabei auf die subjektive Aufgabenattraktivität und aufgabenbezogene Erfolgserwartung fokussiert. Es wurden zudem das Alter und Geschlecht der Schüler:innen berücksichtigt. Der dritte Artikel überprüfte, ob sich Motivthemen in mathematischen Textaufgaben positiv auf die unmittelbare Leistung bei der Aufgabenbearbeitung auswirken. Der vierte Artikel untersuchte den Einfluss motivthematisch formulierter Textaufgaben auf die Leistung Lernender unter Berücksichtigung ihrer individuellen Präferenzen der Lernenden für motivthematisches Material. Hierfür wurden die individuellen Präferenzen für motivthematische Anreizklassen (d.h. Anschluss-, Leistungs- und Machtmotive) der Lernenden ermittelt und zwischen den Hoffnungs- und Furchttendenzen in den Motiven (z.B. Hoffnung auf Erfolg, Furcht vor Misserfolg) differenziert.
Methode. Für die experimentellen Feldstudien (Artikel 1, Artikel 2) wurden Deutsch- (Artikel 1) und Mathematikaufgaben (Artikel 1, Artikel 2) aus gängigen Schulbüchern mit motivthematischen und neutralen Inhalten angereichert. Die Aufgaben wurden Schüler:innen der fünften Jahrgansstufe (Artikel 1) und neunten Jahrgansstufe (Artikel 2) in zufälliger Reihenfolge zur Bewertung in Bezug auf die subjektive Aufgabenattraktivität (Artikel 1, Artikel 2) und auf die aufgabenbezogene Erfolgserwartung (Artikel 1) vorgelegt. In experimentellen Laboruntersuchungen (Artikel 3, Artikel 4), bearbeiteten die Studierenden motivthematisch und neutral formulierte mathematische Textaufgaben, die ihnen in zufälliger Reihenfolge am PC präsentiert wurden. Alle verwendeten Aufgaben waren hinsichtlich Schwierigkeit, Länge und Textkomplexität parallelisiert (Artikel 1-4). Mit Ausnahme der Studie im vierten Artikel, für das ein between-subject Design gewählt wurde, stellten die Motivthemen within-subject Variablen dar (Artikel 1-3). Ergebnisse und Diskussion. Schüler:innen schätzten motivthematisch angereicherte Aufgaben durchgängig als attraktiver ein als neutrale Aufgaben (Artikel 1, Artikel 2) und waren oftmals zuversichtlicher, diese Aufgaben erfolgreich lösen zu können, als motivneutrale Aufgaben (Artikel 1). Die Überlegenheit motivthematischer Aufgaben gegenüber neutralen galt insbesondere für anschlussthematische Aufgaben. Dies war unabhängig vom Geschlecht der Schüler:innen und galt sowohl für Fünftklässler:innen als auch für Neuntklässler:innen. In Bezug auf die Leistung zeigten drei Experimente, dass leistungsthematische Textaufgaben besser gelöst wurden als neutral formulierte Aufgaben (Artikel 3). Für anschlussthematische Aufgaben ergab sich ein uneinheitliches Ergebnismuster, da diese in nur zwei der drei Experimente tendenziell besser gelöst wurden als motivneutrale Aufgaben (Artikel 3). Unter Berücksichtigung der Motive (Artikel 4) wurden differentielle Effekte auf die Leistung in anschlussthematischen Aufgaben in Abhängigkeit der Hoffnungs- und Furchttendenzen im Anschlussmotiv sichtbar. Studierende mit dominierender Hoffnung auf Anschluss erzielten in anschlussthematischen Aufgaben tendenziell bessere Leistungen als in neutralen Aufgaben. Studierende mit überwiegender Furcht vor Zurückweisung waren in anschlussthematischen Aufgaben schlechter als in neutralen Aufgaben. Ein vergleichbares Ergebnismuster für die Leistung in anschlussthematischen Aufgaben zeigte sich darüber hinaus für die Hoffnungs- und Furchttendenzen im Machtmotiv. Die vorliegenden Ergebnisse liefern praktische Hinweise zur Gestaltung von Lernmaterial. So können Anschluss- und Leistungsthemen genutzt werden, um Schulaufgaben für Schüler:innen attraktiver zu gestalten. Zudem wirken sich diese Themen oftmals positiv auf die Leistung aus. Eine differentielle Gestaltung von Schulaufgaben nach den Kriterien Alter und Geschlecht erscheint dabei nicht sinnvoll. Vielmehr legen die vorliegenden Befunde nahe, dass bei der Verwendung anschlussthematischer Inhalte die Hoffnungs- und Furchttendenzen im Anschluss- und Machtmotiv der Lernenden berücksichtigt werden sollten
Schützende Bewältigung. Eine Grounded Theory zu Diskriminierungserfahrungen von Fachkräften in der Sozialen Arbeit.
No full textIn diesem Open-Access-Buch wird eine qualitative Studie zum Thema Diskriminierungserfahrungen von Fachkräften der Sozialen Arbeit vorgestellt. Ziel der Arbeit ist es, eine gegenstandsverankerte Theorie abzubilden, die hilft, den Umgang mit Diskriminierungserfahrung – insbesondere Rassismuserfahrungen – zu beschreiben: Wie verhalten sich Betroffene in diskriminierenden Situationen? Wie gehen sie mit ihren Erfahrungen außerhalb der diskriminierten Situation um? Einer der Schwerpunkte der Untersuchung ist das Zusammenspiel der Erfahrungen und des professionellen Arbeitskontextes. Das Theoriemodell der Schützenden Bewältigung ist ein Beitrag zur rassismuskritischen Sozialen Arbeit und lädt zu einem Perspektivenwechsel ein. Die theoretischen Überlegungen bieten gleichzeitig Anknüpfungspunkte für die Praxis