Refine
Document Type
- Article (6)
- Doctoral Thesis (2)
- Postprint (2)
- Conference Proceeding (1)
Is part of the Bibliography
- yes (11) (remove)
Keywords
- neural networks (11) (remove)
Die trainingswissenschaftliche Diagnostik in den Kernbereichen Training, Wettkampf und Leistungsfähigkeit ist durch einen hohen Praxisbezug, eine ausgeprägte strukturelle Komplexität und vielseitige Wechselwirkungen der sportwissenschaftlichen Teilgebiete geprägt. Diese Eigenschaften haben in der Vergangenheit dazu geführt, dass zentrale Fragestellungen, wie beispielsweise die Maximierung der sportlichen Leistungsfähigkeit, eine ökonomische Trainingsgestaltung, eine effektive Talentauswahl und -sichtung oder die Modellbildung noch nicht vollständig gelöst werden konnten. Neben den bereits vorhandenen linearen Lösungsansätzen werden in dieser Arbeit Methoden aus dem Bereich der Neuronalen Netzwerke eingesetzt. Diese nichtlinearen Diagnoseverfahren sind besonders geeignet für die Analyse von Prozessabläufen, wie sie beispielsweise im Training vorliegen. Im theoretischen Teil werden zunächst Gemeinsamkeiten, Abhängigkeiten und Unterschiede in den Bereichen Training, Wettkampf und Leistungsfähigkeit untersucht sowie die Brücke zwischen trainingswissenschaftlicher Diagnostik und nichtlinearen Verfahren über die Begriffe der Interdisziplinarität und Integrativität geschlagen. Angelehnt an die Theorie der Neuronalen Netze werden anschließend die Grundlagenmodelle Perzeptron, Multilayer-Perzeptron und Selbstorganisierende Karten theoretisch erläutert. Im empirischen Teil stehen dann die nichtlineare Analyse von personalen Anforderungsstrukturen, Zustände der sportlichen Form und die Prognose sportlichen Talents - allesamt bei jugendlichen Leistungsschwimmerinnen und -schwimmern - im Mittelpunkt. Die nichtlinearen Methoden werden dabei einerseits auf ihre wissenschaftliche Aussagekraft überprüft, andererseits untereinander sowie mit linearen Verfahren verglichen.
We present the Neural-network-based Upper hybrid Resonance Determination (NURD) algorithm for automatic inference of the electron number density from plasma wave measurements made on board NASA's Van Allen Probes mission. A feedforward neural network is developed to determine the upper hybrid resonance frequency, fuhr, from electric field measurements, which is then used to calculate the electron number density. In previous missions, the plasma resonance bands were manually identified, and there have been few attempts to do robust, routine automated detections. We describe the design and implementation of the algorithm and perform an initial analysis of the resulting electron number density distribution obtained by applying NURD to 2.5 years of data collected with the Electric and Magnetic Field Instrument Suite and Integrated Science (EMFISIS) instrumentation suite of the Van Allen Probes mission. Densities obtained by NURD are compared to those obtained by another recently developed automated technique and also to an existing empirical plasmasphere and trough density model.
VS30, slope, H800 and f0
(2017)
The aim of this paper is to investigate the ability of various site-condition proxies (SCPs) to reduce ground-motion aleatory variability and evaluate how SCPs capture nonlinearity site effects. The SCPs used here are time-averaged shear-wave velocity in the top 30 m (VS30), the topographical slope (slope), the fundamental resonance frequency (f0) and the depth beyond which Vs exceeds 800 m/s (H800). We considered first the performance of each SCP taken alone and then the combined performance of the 6 SCP pairs [VS30–f0], [VS30–H800], [f0–slope], [H800–slope], [VS30–slope] and [f0–H800]. This analysis is performed using a neural network approach including a random effect applied on a KiK-net subset for derivation of ground-motion prediction equations setting the relationship between various ground-motion parameters such as peak ground acceleration, peak ground velocity and pseudo-spectral acceleration PSA (T), and Mw, RJB, focal depth and SCPs. While the choice of SCP is found to have almost no impact on the median groundmotion prediction, it does impact the level of aleatory uncertainty. VS30 is found to perform the best of single proxies
at short periods (T < 0.6 s), while f0 and H800 perform better at longer periods; considering SCP pairs leads to significant improvements, with particular emphasis on [VS30–H800] and [f0–slope] pairs. The results also indicate significant nonlinearity on the site terms for soft sites and that the most relevant loading parameter for characterising nonlinear site response is the “stiff” spectral ordinate at the considered period.
Increased N400 amplitudes on indefinite articles (a/an) incompatible with expected nouns have been initially taken as strong evidence for probabilistic pre-activation of phonological word forms, and recently been intensely debated because they have been difficult to replicate. Here, these effects are simulated using a neural network model of sentence comprehension that we previously used to simulate a broad range of empirical N400 effects. The model produces the effects when the cue validity of the articles concerning upcoming noun meaning in the learning environment is high, but fails to produce the effects when the cue validity of the articles is low due to adjectives presented between articles and nouns during training. These simulations provide insight into one of the factors potentially contributing to the small size of the effects in empirical studies and generate predictions for cross-linguistic differences in article induced N400 effects based on articles’ cue validity. The model accounts for article induced N400 effects without assuming pre-activation of word forms, and instead simulates these effects as the stimulus-induced change in a probabilistic representation of meaning corresponding to an implicit semantic prediction error.
Increased N400 amplitudes on indefinite articles (a/an) incompatible with expected nouns have been initially taken as strong evidence for probabilistic pre-activation of phonological word forms, and recently been intensely debated because they have been difficult to replicate. Here, these effects are simulated using a neural network model of sentence comprehension that we previously used to simulate a broad range of empirical N400 effects. The model produces the effects when the cue validity of the articles concerning upcoming noun meaning in the learning environment is high, but fails to produce the effects when the cue validity of the articles is low due to adjectives presented between articles and nouns during training. These simulations provide insight into one of the factors potentially contributing to the small size of the effects in empirical studies and generate predictions for cross-linguistic differences in article induced N400 effects based on articles’ cue validity. The model accounts for article induced N400 effects without assuming pre-activation of word forms, and instead simulates these effects as the stimulus-induced change in a probabilistic representation of meaning corresponding to an implicit semantic prediction error.
We argue that natural language can be usefully described as quasi-compositional and we suggest that deep learning-based neural language models bear long-term promise to capture how language conveys meaning. We also note that a successful account of human language processing should explain both the outcome of the comprehension process and the continuous internal processes underlying this performance. These points motivate our discussion of a neural network model of sentence comprehension, the Sentence Gestalt model, which we have used to account for the N400 component of the event-related brain potential (ERP), which tracks meaning processing as it happens in real time. The model, which shares features with recent deep learning-based language models, simulates N400 amplitude as the automatic update of a probabilistic representation of the situation or event described by the sentence, corresponding to a temporal difference learning signal at the level of meaning. We suggest that this process happens relatively automatically, and that sometimes a more-controlled attention-dependent process is necessary for successful comprehension, which may be reflected in the subsequent P600 ERP component. We relate this account to current deep learning models as well as classic linguistic theory, and use it to illustrate a domain general perspective on some specific linguistic operations postulated based on compositional analyses of natural language. This article is part of the theme issue 'Towards mechanistic models of meaning composition'.
Increased N400 amplitudes on indefinite articles (a/an) incompatible with expected nouns have been initially taken as strong evidence for probabilistic pre-activation of phonological word forms, and recently been intensely debated because they have been difficult to replicate. Here, these effects are simulated using a neural network model of sentence comprehension that we previously used to simulate a broad range of empirical N400 effects. The model produces the effects when the cue validity of the articles concerning upcoming noun meaning in the learning environment is high, but fails to produce the effects when the cue validity of the articles is low due to adjectives presented between articles and nouns during training. These simulations provide insight into one of the factors potentially contributing to the small size of the effects in empirical studies and generate predictions for cross-linguistic differences in article induced N400 effects based on articles’ cue validity. The model accounts for article induced N400 effects without assuming pre-activation of word forms, and instead simulates these effects as the stimulus-induced change in a probabilistic representation of meaning corresponding to an implicit semantic prediction error.
Supervised machine learning to assess methane emissions of a dairy building with natural ventilation
(2020)
A reliable quantification of greenhouse gas emissions is a basis for the development of adequate mitigation measures. Protocols for emission measurements and data analysis approaches to extrapolate to accurate annual emission values are a substantial prerequisite in this context. We systematically analyzed the benefit of supervised machine learning methods to project methane emissions from a naturally ventilated cattle building with a concrete solid floor and manure scraper located in Northern Germany. We took into account approximately 40 weeks of hourly emission measurements and compared model predictions using eight regression approaches, 27 different sampling scenarios and four measures of model accuracy. Data normalization was applied based on median and quartile range. A correlation analysis was performed to evaluate the influence of individual features. This indicated only a very weak linear relation between the methane emission and features that are typically used to predict methane emission values of naturally ventilated barns. It further highlighted the added value of including day-time and squared ambient temperature as features. The error of the predicted emission values was in general below 10%. The results from Gaussian processes, ordinary multilinear regression and neural networks were least robust. More robust results were obtained with multilinear regression with regularization, support vector machines and particularly the ensemble methods gradient boosting and random forest. The latter had the added value to be rather insensitive against the normalization procedure. In the case of multilinear regression, also the removal of not significantly linearly related variables (i.e., keeping only the day-time component) led to robust modeling results. We concluded that measurement protocols with 7 days and six measurement periods can be considered sufficient to model methane emissions from the dairy barn with solid floor with manure scraper, particularly when periods are distributed over the year with a preference for transition periods. Features should be normalized according to median and quartile range and must be carefully selected depending on the modeling approach.
We consider a ring network of theta neurons with non-local homogeneous coupling. We analyse the corresponding continuum evolution equation, analytically describing all possible steady states and their stability. By considering a number of different parameter sets, we determine the typical bifurcation scenarios of the network, and put on a rigorous footing some previously observed numerical results.
Faced with the triad of time-cost-quality, the realization of production tasks under economic conditions is not trivial. Since the number of Artificial-Intelligence-(AI)-based applications in business processes is increasing more and more nowadays, the efficient design of AI cases for production processes as well as their target-oriented improvement is essential, so that production outcomes satisfy high quality criteria and economic requirements. Both challenge production management and data scientists, aiming to assign ideal manifestations of artificial neural networks (ANNs) to a certain task. Faced with new attempts of ANN-based production process improvements [8], this paper continues research about the optimal creation, provision and utilization of ANNs. Moreover, it presents a mechanism for AI case-based reasoning for ANNs. Experiments clarify continuously improving ANN knowledge bases by this mechanism empirically. Its proof-of-concept is demonstrated by the example of four production simulation scenarios, which cover the most relevant use cases and will be the basis for examining AI cases on a quantitative level.