489 research outputs found
Using autoencoders on differentially private federated learning GANs
No full textMachine learning has been applied to almost all fields of computer science over the past decades. The introduction of GANs allowed for new possibilities in fields of medical research and text prediction. However, these new fields work with ever more privacy-sensitive data. In order to maintain user privacy, a combination of federated learning, differential privacy and GANs can be used to work with private data without giving away a users' privacy. Recently, two implementations of such combinations have been published: DP-Fed-Avg GAN and GS-WGAN. This paper compares their performance and introduces an alternative version of DP-Fed-Avg GAN that makes use of denoising techniques to combat the loss in accuracy that generally occurs when applying differential privacy and federated learning to GANs. We also compare the novel adaptation of denoised DP-Fed-Avg GAN to the state-of-the-art implementations in this field.CSE3000 Research ProjectComputer Science and Engineerin
Monte Carlo simulation of SDEs using GANs
No full textGenerative adversarial networks (GANs) have shown promising results when applied on partial differential equations and financial time series generation. We investigate if GANs can also be used to approximate one-dimensional Ito ^ stochastic differential equations (SDEs). We propose a scheme that approximates the path-wise conditional distribution of SDEs for large time steps. Standard GANs are only able to approximate processes in distribution, yielding a weak approximation to the SDE. A conditional GAN architecture is proposed that enables strong approximation. We inform the discriminator of this GAN with the map between the prior input to the generator and the corresponding output samples, i.e. we introduce a ‘supervised GAN’. We compare the input-output map obtained with the standard GAN and supervised GAN and show experimentally that the standard GAN may fail to provide a path-wise approximation. The GAN is trained on a dataset obtained with exact simulation. The architecture was tested on geometric Brownian motion (GBM) and the Cox–Ingersoll–Ross (CIR) process. The supervised GAN outperformed the Euler and Milstein schemes in strong error on a discretisation with large time steps. It also outperformed the standard conditional GAN when approximating the conditional distribution. We also demonstrate how standard GANs may give rise to non-parsimonious input-output maps that are sensitive to perturbations, which motivates the need for constraints and regularisation on GAN generators.ImPhys/Practicum supportNumerical Analysi
Correspondence to Albert Paul Brinson from Curtis B. Gans, May 2, 1960
No full textThis document contains a letter addressed to Mr. Albert Paul Brinson from Curtis B. Gans, the National Affairs Vice President of the United States National Student Association. The letter expresses gratitude for Mr. Brinson's participation in the National Student Conference on the Sit-In Movement, held in Washington. Gans commends Brinson's courage and leadership in the civil rights movement, highlighting that his contribution inspires others. The letter emphasizes the importance of building a nation where all individuals can experience freedom. It concludes with appreciation for Brinson's efforts and encourages him to continue the fight for positive change. 1 page
Genève, La Bibliothèque juive « Gérard Nordmann », HEB 0002 : Astronomical and geographical work
No full textThis manuscript is a copy of only one of six extant manuscript exemplars and an old print of this work (1743) worldwide. Its author is the famous Bohemian rabbi, astronomer and mathematician R. David ben Salomon Gans (1541-1613). In his 1974 monograph about David Gans, André Neher referred to this copy as the Manuscrit de Genève. A colophon in the manuscript gives the date as 1613, but a current study on the history of the transmission of this work suggests that it is an 18th century copy.Online Since: 2018-06-1
Generating Asset Paths for Financial SDEs with GANs
No full textGenerative adversarial networks (GANs) have shown promising results when applied on partial differential equations and financial time series generation. This thesis investigates if GANs can be used to provide a strong approximation to the solution of stochastic differential equations (SDEs) of the Ito type. Standard GANs are only able to approximate processes in conditional distribution, yielding a weak approximation to the SDE. A novel GAN architecture is proposed that enables strong approximation, called the constrained GAN. The discriminator of this GAN is informed with the random sample that corresponds to the Brownian motion increment between two time steps. This way, the constrained GAN does not only learn the conditional distribution, but the unique map from a random increment to the next asset value along the path, conditional on the previous value. The architecture was tested on geometric Brownian motion (GBM) and the Cox Ingersoll Ross (CIR) process in one dimension, where it was conditioned on a range of time steps and previous values of the asset process. The constrained GAN was shown to outperform discrete-time schemes in strong error on a discretisation with large time steps. It also outperformed the standard conditional GAN when approximating the conditional distribution. A method is proposed to extend the constrained GAN to general one-dimensional Ito SDEs, beyond the SDEs tested in this work. In future work, the constrained GAN should be conditioned on the SDE parameters as well, allowing it to learn an entire family of solutions at once. Furthermore, the architecture could be extended to higher dimensions, including systems of SDEs, such as the Heston model.Applied Mathematics | Financial Engineerin
Eduard Gans y la idea de Europa
No full textEl presente artículo pretende ser una aproximación a la concepción de Europa de Eduard Gans, un autor al que un número creciente de intérpretes considera como el que es, quizá, el discípulo más destacado de Hegel. El tema de Europa era entonces de gran actualidad debido tanto al influjo de la Ilustración y la Revolución francesa como al desarrollo del Idealismo alemán. Gans está en estrecha relación con estas instancias y formula una idea de Europa que termina por ir más allá de la de su maestro Hegel. Se abre a los nuevos horizontes que se van configurando a partir de 1830 en Europa, a la vez que se muestra receptivo a los estímulos provenientes de América.This article is supposed to be an approximation to Eduard Gans´ conception of Europe, an author considered to be the most prominent disciple of Hegel by a growing number of scholars. In those times, the idea of Europe was a highly topical subject, due to both to the influence of the Enlightenment and the French Revolution, but as well to the development of German idealism. Gans is closely related to these instances and formulates an idea of Europe that goes beyond the conception of his teacher Hegel. He takes into account the new instances that arise from 1830 on in Europe, while he is also receptive to the views coming from America
Time series synthesis using GANs - A take on DoppelGANger
No full textWith a growing need for data comes a growing need for synthetic data. In this work we reproduce the results of DoppelGANger [16] in synthesising time series data with metadata. We identify a key issue in the comparison made in [16] of DoppelGANger to TimeGAN, RNNs, AR and HMM models, which creates a new avenue of time series synthesis using GANs. We show that not all results of [16] can be reproduced. We furthermore find that DoppelGANger does not adequately capture measurement-metadata correlations of our dataset. Sample size reduction is shown to be an effective tool to reduce training time while still attaining accurate results, and the key parameter S is tuned further. Finally we show that execution on CPU has similar training times as execution on GPU by [16], suggesting that the original code can be improved, and we release our version of the models ourselves, to enable easy reproduction. In closing points we shine light on possible future improvements that we were unable to test ourselves, and conclude that DoppelGANger is a promising model that opens the door to new unseen applications of GANs for time series synthesis.CSE3000 Research ProjectComputer Science and Engineerin
The Benefits of Being Economics Professor A (and not Z)
No full textAlphabetic name ordering on multi-authored academic papers, which is the convention in the economics discipline and various other disciplines, is to the advantage of people whose last name initials are placed early in the alphabet. As it turns out, Professor A, who has been a first author more often than Professor Z, will have published more articles and experienced afaster growth rate over the course of her career as a result of reputation and visibility. Moreover, authors know that name ordering matters and indeed take ordering seriously: Several characteristics of an author group composition determine the decision to deviate from the default alphabetic name order to a significant extent.performance measurement, incentives, economists, name ordering
GDTS: GAN-Based Distributed Tabular Synthesizer
No full textGenerative Adversarial Networks (GANs) are typically trained to synthesize data, from images and more recently tabular data, under the assumption of directly accessible training data. While learning image GANs on Federated Learning (FL) and Multi-Discriminator (MD) systems has just been demonstrated, it is unknown if tabular GANs can be learned from decentralized data sources. Different from image GANs, state-of-the-art tabular GANs require prior knowledge on the data distribution of each (discrete and continuous) column to agree on a common encoding - risking privacy guarantees. In this paper, we propose GDTS, a distributed framework for GAN-based tabular synthesizer. GDTS provides different system architectures to match the two training paradigms termed GDTS_FL and GDTS_MD. Key to enable learning on distributed data is the proposed novel privacy-preserving multi-source feature encoding to capture the global data properties. In addition GDTS encompasses a weighting strategy based on table similarity to counter the detrimental effects of non-IID data and a validation pipeline to easily assess and compare the performance of different paradigms and hyper parameters. We evaluate the effectiveness of GDTS in terms of synthetic data quality, and overall training scalability. Experiments show that GDTS_FL achieves better statistical similarity and machine learning utility between generated and original data compared to GDTS_MD.Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.Data-Intensive System
The Augmented Dataset: Artistic Appropriations of GANs and their Bearings on Ethical Considerations of Artificial Intelligence
No full textInternational audienceGenerative Adversarial Networks (GANs) have received much recent attention as they have been employed to falsify information through various media channels and to visually mislead viewers in their interpretation of still images and video. Decried under the rubric of fake news, GANs are often held up as malefactors in the crusade against unethical AI, yet their applications are wide ranging and their potential has yet to be fully realized. This presentation investigates the use of GANs by artists as an alternative to this narrative and considers the role of dataset formation in the Artificial Intelligence artistic process. Since such a significant number of images is required for machine learning systems to function well, the need to augment a dataset is often encountered and how this is overcome plays a considerable role in the final visual form of the GANs-produced image. Indeed, artistic approaches to the hurdle to create new digital images through a repository of so many existing ones offer insights on what constitutes ethical Artificial Intelligence practices. The examples considered include those by notable Artificial Intelligence artists along with a recent project on gastronomic algorithms undertaken by the author with the Chef Alain Passard. Together, these artworks and projects lead one to the question: How can an image that is created through its computational yet obscured connection to a plethora of images be measured at all
- …
