scholarly journals Better latent spaces for better autoencoders

2021 ◽  
Vol 11 (3) ◽  
Author(s):  
Barry Dillon ◽  
Tilman Plehn ◽  
Christof Sauer ◽  
Peter Sorrenson
Keyword(s):  
Gaussian Mixture ◽  
The Other ◽  
Structural Problem ◽  
Latent Space

Autoencoders as tools behind anomaly searches at the LHC have the structural problem that they only work in one direction, extracting jets with higher complexity but not the other way around. To address this, we derive classifiers from the latent space of (variational) autoencoders, specifically in Gaussian mixture and Dirichlet latent spaces. In particular, the Dirichlet setup solves the problem and improves both the performance and the interpretability of the networks.

The Dark Tetrad

2015 ◽  
Vol 36 (4) ◽  
pp. 228-236 ◽  
Author(s):  
Janko Međedović ◽  
Boban Petrović
Keyword(s):  
Personality Traits ◽  
Antisocial Behavior ◽  
Second Order ◽  
The Other ◽  
Negative Pole ◽  
Latent Space ◽  
Order Factor

Abstract. Machiavellianism, narcissism, and psychopathy are personality traits understood to be dispositions toward amoral and antisocial behavior. Recent research has suggested that sadism should also be added to this set of traits. In the present study, we tested a hypothesis proposing that these four traits are expressions of one superordinate construct: The Dark Tetrad. Exploration of the latent space of four “dark” traits suggested that the singular second-order factor which represents the Dark Tetrad can be extracted. Analysis has shown that Dark Tetrad traits can be located in the space of basic personality traits, especially on the negative pole of the Honesty-Humility, Agreeableness, Conscientiousness, and Emotionality dimensions. We conclude that sadism behaves in a similar manner as the other dark traits, but it cannot be reduced to them. The results support the concept of “Dark Tetrad.”

scholarly journals Variational Information Bottleneck for Unsupervised Clustering: Deep Gaussian Mixture Embedding

Entropy ◽  
2020 ◽  
Vol 22 (2) ◽  
pp. 213 ◽  
Author(s):  
Yiğit Uğur ◽  
George Arvanitakis ◽  
Abdellatif Zaidi
Keyword(s):  
Neural Networks ◽  
Lower Bound ◽  
Gradient Descent ◽  
Gaussian Mixture ◽  
Variational Inference ◽  
Stochastic Gradient Descent ◽  
Information Bottleneck ◽  
Latent Space ◽  
Type Algorithm ◽  
The Cost

In this paper, we develop an unsupervised generative clustering framework that combines the variational information bottleneck and the Gaussian mixture model. Specifically, in our approach, we use the variational information bottleneck method and model the latent space as a mixture of Gaussians. We derive a bound on the cost function of our model that generalizes the Evidence Lower Bound (ELBO) and provide a variational inference type algorithm that allows computing it. In the algorithm, the coders’ mappings are parametrized using neural networks, and the bound is approximated by Markov sampling and optimized with stochastic gradient descent. Numerical results on real datasets are provided to support the efficiency of our method.

Introduction

2019 ◽  
pp. 1-12
Author(s):  
Eva Meijer
Keyword(s):  
San Diego ◽  
The Other ◽  
Structural Problem ◽  
Hands On ◽  
Break Up

Orangutan Ken Allen was born in the San Diego Zoo. While still in the nursery he was already trying to unscrew every nut he could get his hands on, and he used humans as objects to climb on in order to escape the room. In the years that followed, he perfected his techniques, which led him to escape his enclosure many times. This forced the zoo to alter the fences around the orangutan enclosure and change their windows and locks. They also tried to distract him by bringing in females, and they hired spies posing as visitors in an attempt to find out how he did it. But Ken Allen was not the only orangutan who had a desire to leave captivity. His mate Vicki once took over from him, unbolting a door after he was caught, and escaped. Kumang and Sara, two sisters from the same group, organized and coordinated their own escapes, for example, by using a mop handle that one of them held in place while the other climbed it. Cooperative orangutan resistance is found in many other zoos as well, including in the Woodland Park Zoo in Seattle, where a group of five orangutans slipped through several security doors and climbed over a high wall. Neither bananas nor water from fire hoses could convince them to go back in, and they had to be tranquilized. These examples are not the only ones available—orangutan resistance is a structural problem for zoos, often leading them to isolate individuals or break up family bonds by relocating orangutans (...

Vehicle Appearance Model for Recognition System Considering the Change of Imaging Condition

2009 ◽  
Vol 13 (4) ◽  
pp. 463-469 ◽  
Author(s):  
Keiji Kuwabara ◽  
◽  
Yoshikazu Yano ◽  
Shigeru Okuma ◽  
Keyword(s):  
Gaussian Mixture Model ◽  
Mixture Model ◽  
Gaussian Mixture ◽  
Recognition System ◽  
The Other ◽  
Appearance Model ◽  
Imaging Condition ◽  
Successive Change ◽  
Imaging Conditions ◽  
And Training

We have proposed a technique to recognize a vehicle. In this technique, Gaussian Mixture Model (GMM) is adopted as a classifier. Vehicle appearance changed by imaging conditions such as time, weather and so on, and GMM parameters are also changed by imaging conditions. To recognize vehicle accurately, we have prepared some GMM tuned with the imaging conditions. On the other hand, it is impossible to prepare GMM because imaging condition changes successively. In this paper, we propose a method for estimating GMM and for training GMM parameters which reflect the successive change of imaging condition. Experimental results show that GMM parameters are estimated accurately and training of GMM are speeded up by proposed method.

Semi-supervised fabric defect detection based on image reconstruction and density estimation

2020 ◽  
pp. 004051752096673
Author(s):  
Qihong Zhou ◽  
Jun Mei ◽  
Qian Zhang ◽  
Shaozong Wang ◽  
Ge Chen
Keyword(s):  
Image Reconstruction ◽  
Density Estimation ◽  
Defect Detection ◽  
Gaussian Mixture ◽  
Inspection System ◽  
Fabric Defect Detection ◽  
Latent Space ◽  
Variational Autoencoder ◽  
Fabric Defect ◽  
Fabric Surface

Defective products are a major contributor toward a decline in profits in textile industries. Hence, there are compelling needs for an automated inspection system to identify and locate defects on the fabric surface. Although much effort has been made by researchers worldwide, there are still challenges with computation and accuracy in the location of defects. In this paper, we propose a hybrid semi-supervised method for fabric defect detection based on variational autoencoder (VAE) and Gaussian mixture model (GMM). The VAE model is trained for feature extraction and image reconstruction while the GMM is used to perform density estimation. By synthesizing the detection results from both image content and latent space, the method can construct defect region boundaries more accurately, which are useful in fabric quality evaluation. The proposed method is validated on AITEX and DAGM 2007 public database. Results demonstrate that the method is qualified for automated detection and outperforms other selected methods in terms of overall performance.

High Hopes, Scant Resources: A Word of Scepticism about the Anti-Fragmentation Function of Article 31(3)(c) of the Vienna Convention on the Law of Treaties

2011 ◽  
Vol 24 (3) ◽  
pp. 701-714 ◽  
Author(s):  
MÉLANIE SAMSON
Keyword(s):  
Present Article ◽  
Fragmentation Function ◽  
The Other ◽  
Vienna Convention ◽  
Structural Problem ◽  
Other Hand ◽  
Interpretive Approach ◽  
The Law

AbstractFocusing on some undertheorized aspects of Article 31(3)(c) of the Vienna Convention on the Law of Treaties, the present article aims to reassess critically the anti-fragmentation function generally assigned to this provision. The high hopes associated with the harmonizing potential of Article 31(3)(c) are usually based on a reading of this provision as requiring the interpreter to take into account not only rules applicable between all of the parties to the treaty, but also those applicable only between some of the parties. However, this reading does not seem to be confirmed by the interpretive approach suggested in this article. On the other hand, the use of Article 31(3)(c) in judicial settings raises a structural problem inherent in the international judiciary. The analysis undertaken along these lines suggests that the optimism that Article 31(3)(c) has recently provoked should be qualified in some important respects.

scholarly journals Semi-supervised Adversarial Variational Autoencoder

2020 ◽  
Author(s):  
Ryad Zemouri
Keyword(s):  
Image Processing ◽  
Learning Process ◽  
Original Data ◽  
Experimental Results ◽  
The Other ◽  
Training Process ◽  
Adversarial Learning ◽  
Latent Space ◽  
Variational Autoencoder ◽  
Label Information

We present a method to improve the reconstruction and generation performance of variational autoencoder (VAE) by injecting an adversarial learning. On the other hand, instead of comparing the reconstructed with the original data to calculate the reconstruction loss, we use a consistency principle for deep features. The training process of the VAE is then divided into two steps, training the encoder and then training the decoder. By using this two-step learning process, our method can be more widely used in applications other than image processing. While training the encoder, the label information is integrated to better structure the latent space in a supervised way. The adversarial constraints allow the decoder to generate data with better authenticity and more realistic than the conventional VAE. We present experimental results to show that our method gives better performance than the original VAE.

scholarly journals Tumor Detection in Brain MRIs by Computing Dissimilarities in the Latent Space of a Variational AutoEncoder

2020 ◽  
Vol 1 ◽  
pp. 6
Author(s):  
Alexandra Albu ◽  
Alina Enescu ◽  
Luigi Malagò
Keyword(s):  
Detection Method ◽  
Brain Mri ◽  
The Other ◽  
Detection Methods ◽  
Healthy Individuals ◽  
Latent Space ◽  
Variational Autoencoder ◽  
Mri Scans ◽  
Unsupervised Anomaly Detection ◽  
Aided Diagnosis

The ability to automatically detect anomalies in brain MRI scans is of great importance in computer-aided diagnosis. Unsupervised anomaly detection methods work primarily by learning the distribution of healthy images and identifying abnormal tissues as outliers. We propose a slice-wise detection method which first trains a pair of autoencoders on two different datasets, one with healthy individuals and the other one with images of normal and tumoral tissues. Next, it classifies slices based on the distance in the latent space between the enconding of the image and the encoding of the reconstructed image, obtained through the autoencoder trained on healthy images only. We validate our approach with a series of preliminary experiments on the HCP and BRATS-15 datasets.

scholarly journals Combining outlier analysis algorithms to identify new physics at the LHC

2021 ◽  
Vol 2021 (9) ◽  
Author(s):  
Melissa van Beekveld ◽  
Sascha Caron ◽  
Luc Hendriks ◽  
Paul Jackson ◽  
Adam Leinweber ◽  
...  
Keyword(s):  
Hadron Collider ◽  
Gaussian Mixture ◽  
New Physics ◽  
Outlier Analysis ◽  
Detection Techniques ◽  
Latent Space ◽  
Variational Autoencoder ◽  
Model Independent ◽  
Weighted Combination ◽  
Isolation Forest

Abstract The lack of evidence for new physics at the Large Hadron Collider so far has prompted the development of model-independent search techniques. In this study, we compare the anomaly scores of a variety of anomaly detection techniques: an isolation forest, a Gaussian mixture model, a static autoencoder, and a β-variational autoencoder (VAE), where we define the reconstruction loss of the latter as a weighted combination of regression and classification terms. We apply these algorithms to the 4-vectors of simulated LHC data, but also investigate the performance when the non-VAE algorithms are applied to the latent space variables created by the VAE. In addition, we assess the performance when the anomaly scores of these algorithms are combined in various ways. Using super- symmetric benchmark points, we find that the logical AND combination of the anomaly scores yielded from algorithms trained in the latent space of the VAE is the most effective discriminator of all methods tested.

scholarly journals From Galileo to Navier and Clapeyron

2021 ◽  
Vol 10 (2) ◽  
pp. 5-20
Author(s):  
Josep Maria Pons
Keyword(s):  
History Of Science ◽  
The Other ◽  
Breaking Point ◽  
Strength Of Materials ◽  
Structural Problem ◽  
Learning And Teaching ◽  
Cooperative Approach ◽  
Present Essay ◽  
History Of ◽  
Interesting Discussion

Galileo (1564-1642), in his well-known Discorsi (Galileo, 1638), briefly turning his attention to the fracture of a beam, starts an interesting discussion on the beam’s breakage as well as its location. Could the section and breaking point of a beam have been determined beforehand? Furthermore, is it specific to the material? What Galileo did was not merely challenge a physics problem, but the prevailing knowledge of his time: namely, Aristotelianism on one hand, and Nominalism on the other. As a matter of fact, must the breakage of an element be treated as a universal or is it particular to a given material? The present essay aims to prove how Galileo, confronting the structural problem and bringing it into the realm of science, was not just raising a problem but, using Salviati’s words, he also established what actually takes place. Many years later, with the progress of physics, strength of materials and theory of structures, figures such as Claude Navier (1785-1836) and Benoît Clapeyron (1799-1864) confirmed once again that the Pisan turned out to be right. This article intends to combine technical fields such as strength of materials and theory of structures with others like the history of science and philosophy proper. A cooperative approach to these disciplines can be doubtlessly helpful to improve the knowledge, learning and teaching of their different curricula, giving the reader a global, holistic perspective.  

Sign in / Sign up

Export Citation Format

Share Document