4,739 research outputs found
Behavioral modeling of GaN-based power amplifiers: impact of electrothermal feedback on the model accuracy and identification
No full textIn this article, we discuss the accuracy of behavioral models in simulating the intermodulation distortion (IMD) of microwave GaN-based high-power amplifiers in the presence of strong electrothermal (ET) feedback. Exploiting an accurate self-consistent ET model derived from measurements and thermal finite-element method simulations, we show that behavioral models are able to yield accurate results, provided that the model identification is carried out with signals with wide bandwidth and large dynamics
LCT-GAN - Improving the efficiency of tabular data synthesis via latent embeddings
No full textIn the past decade data-driven approaches have been at the core of many business and research models. In critical domains such as healthcare and banking, data privacy issues are very stringent. Synthetic tabular data is an emerging solution to privacy guarantee concerns. Generative Adversarial Networks (GANs) are one of the emerging solutions for synthesizing data. However in order to capture all relevant relationships between columns, tabular data needs to be numerically encoded. As columns might be of different types, this is a challenging task as proven by recent approaches. Throughout this paper, we focus on the dimensionality explosion problem, which leads to high-dimensional datasets alongside computational overhead and increase in training time. We introduce a novel synthesis pipeline - LCT-GAN - an improvement to the current state-of-the-art in tabular data synthesis CTAB-GAN. Our approach addresses the dimensionality explosion problem by introducing a low-dimensional embedding step via an autoencoder prior to training. It is then combined with a novel conditional GAN architecture, operating in latent space. After thorough evaluation, we observe that our solution achieves more than 30\% improvement in certain statistical metrics in comparsion to CTAB-GAN, accompanied by 5 fold decrease in size and 150 times speedup in training time for a single epoch. We successfully show that it is possible to embed data using autoencoders, and that GANs are able to learn complex relationships in latent space in the context of tabular data.CSE3000 Research ProjectComputer Science and Engineerin
FCT-GAN: Fourier Neural Operator for Global Relation Enhancement in Tabular Data Synthesizing using Generative Adversarial Networks
No full textSince the regularization of data privacy (e.g., GDPR), the effectiveness of data sharing has decreased. A promising technique to circumvent this problem is tabular data synthesis (i.e., the generation of fake tabular data that statistically resembles the original data). However, the state-of-the-art tabular data synthesis model, CTAB-GAN, fails at robustly imitating global data dependencies and underperforms when column orders get permuted. CTAB-GAN internally uses Convolutional Neural Networks (CNN) which limits the model’s performance due to a strictly non-global data perspective during iterative training phases. To address this limitation, this paper proposes FCT-GAN which leverages the Fourier Neural Operator to learn global dependencies in the frequency domain. Specifically, it enhances CTAB-GAN by replacing the CNN of the discriminator with a four-layered two-dimensional Fourier Neural Operator. As a consequence of FCT-GAN’s global nature and cross-column relation robustness, it outperforms CTAB-GAN and additionally offers the column permutation invariant property. The evaluation of FCT-GAN on five datasets shows that the generated data, remarkably resembles the real data and reveals an increase in accuracy, by up to 19% for five machine learning algorithms independent of data column order, compared to CTAB-GAN.CSE3000 Research ProjectComputer Science and Engineerin
Band offsets of non-polar A-plane GaN/AlN and AlN/GaN heterostructures measured by X-ray photoemission spectroscopy
No full textThe band offsets of non-polar A-plane GaN/AlN and AlN/GaN heterojunctions are measured by X-ray photoemission spectroscopy. A large forward-backward asymmetry is observed in the non-polar GaN/AlN and AlN/GaN heterojunctions. The valence-band offsets in the non-polar A-plane GaN/AlN and AlN/GaN heterojunctions are determined to be 1.33 +/- 0.16 and 0.73 +/- 0.16 eV, respectively. The large valence-band offset difference of 0.6 eV between the non-polar GaN/AlN and AlN/GaN heterojunctions is considered to be due to piezoelectric strain effect in the non-polar heterojunction overlayers.Nanoscience & NanotechnologyMaterials Science, MultidisciplinaryPhysics, AppliedSCI(E)[email protected]; [email protected]
The influence of the 1st AlN and the 2nd GaN layers on properties of AlGaN/2nd AlN/2nd GaN/1st AlN/1st GaN structure
No full textThis is a theoretical study of the 1st AlN interlayer and the 2nd GaN layer on properties of the Al(0.3)Ga(0.7)N/2nd AlN/2nd GaN/1st AlN/1st GaN HEMT structure by self-consistently solving coupled Schrodinger and Poisson equations. Our calculation shows that by increasing the 1st AlN thickness from 1.0 nm to 3.0 nm, the 2DEG, which is originally confined totally in the 2nd channel, gradually decreases, begins to turn up and eventually concentrates in the 1st one. The total 2DEG (2DEG in both channels) sheet density increases nearly linearly with the increasing 1st AlN thickness. And the slope of the potential profile of the AlGaN changes with the 1st AlN thickness, causing the unusual dependence of the total 2DEG sheet density on the thickness of the AlGaN barrier. The variations of 2DEG distribution, the total 2DEG sheet density and the conduction band profiles as a function of the 2nd GaN thickness also have been discussed. Their physical mechanisms have been investigated on the basis of the surface state theory. And the confinement of 2DEG can be further enhanced by the double-AlN interlayer, compared with the InGaN back-barrier
Author reply: Wei Gan JJ, Lia Gan JJ, Hsien Gan JJ, Lee KT. Lateral percutaneous nephrolithotomy: A safe and effective surgical approach. Indian J Urol 2018;34:45-50
No full textLow power AlGaN/GaN MEMS pressure sensor for high vacuum application
No full textA micro-scale pressure sensor based on suspended AlGaN/GaN heterostructure is reported with non-linear sensitivity. By sealing the cavity, vacuum sensing at various temperatures was demonstrated. To validate the proposed concept of the AlGaN/GaN vacuum sensor, a 700 µm diameter circular membrane was electrically characterized under applied static and dynamic pressures at various temperatures ranging from 25 °C to 100 °C. The current change of the AlGaN/GaN heterostructure increased as the vacuum and temperature increases due to the increase of 2DEG density by tensile strain. The dynamic current change from 96 kPa down to 10 Pa of AlGaN/GaN heterostructure pressure sensor was 18.75 % at 100 °C. The maximum sensitivity reached 22.8 %/kPa with a power consumption of 1.8 µW. These results suggest that suspended AlGaN/GaN heterostructures are promising for high vacuum and high-temperature sensing applications.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.Electronic Components, Technology and Material
Extending CTAB-GAN+ with StackGAN architecture
No full textAs privacy regulations (e.g. European General Data Protection Regulation) often prevent valuable flows of data between stakeholders, data synthesis can play a crucial role in sharing captured value in data sets without sharing personal details. Different attempts have been made at solving this problem with Generative Adversarial Networks (GAN) architectures. The aim of this study to create an architecture that can create data that is more resemblant of an original dataset than what the state-of-the-art can achieve, while functioning under the same privacy budget. We apply the concept of progressive learning on the state of the art tabular GAN, i.e., stacking two CTAB-GAN in a sequential manner. Two CTAB-GAN architectures were used, where Stage I CTAB-GAN is trained as usual and the Stage II CTAB-GAN is trained using the result of Stage I CTAB-GAN and the same conditional vector that was used for Stage I. Stage II CTABGAN can rectify defects made in Stage I and further refine the the results to finally achieve a generative model that can better capture the information in the original dataset. The challenge lies in creating the condition in which the second GAN can actually improve on the first GANs results. The results on four datasets show that better data synthesis can be achieved with a second GAN that has a fully connected generator, as opposed to a copy of the original generator of CTAB-GAN.Computer Science and Engineerin
Suppression of persistent photoconductivity AlGaN/GaN heterostructure photodetectors using pulsed heating
No full textThis paper demonstrates a method to reduce the decay time in AlGaN/GaN photodetectors by a pulsed heating mode. A suspended AlGaN/GaN heterostructure photodetector integrated with a micro-heater is fabricated and characterized under ultraviolet illumination. We have observed that the course of persistent photoconductivity was effectively accelerated by applying pulsed heating. The decay time is significantly reduced from 175 s by DC heating to 116 s by 50 Hz pulsed heating at the same power (280 mW). With the same pulse duty cycle and a 50 Hz pulsed heating frequency, a reduction of 30%-45% in decay time is measured compared to DC heating.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.Electronic Components, Technology and Material
Forged-GAN-BERT: Authorship Attribution for LLM-Generated Forged Novels
No full textThe advancement of generative Large Language Models (LLMs), capable of producing human-like texts, introduces challenges related to the authenticity of the text documents. This requires exploring potential forgery scenarios within the context of authorship attribution, especially in the literary domain. Particularly, two aspects of doubted authorship may arise in novels, as a novel may be imposed by a renowned author or include a copied writing style of a well-known novel. To address these concerns, we introduce Forged-GAN-BERT, a modified GAN-BERT-based model to improve the classification of forged novels in two data-augmentation aspects: via the Forged Novels Generator (i.e., ChatGPT) and the generator in GAN. Compared to other transformer-based models, the proposed Forged-GAN-BERT model demonstrates an improved performance with F1 scores of 0.97 and 0.71 for identifying forged novels in single-author and multi-author classification settings. Additionally, we explore different prompt categories for generating the forged novels to analyse the quality of the generated texts using different similarity distance measures , including ROUGE-1, Jaccard Similarity, Overlap Confident, and Cosine Similarity
- …
