arXiv subjects

Jacek Tabor

Publications and source records attributed to Jacek Tabor.

At least 109 records · Page 6Linked to original sources

arXiv2020

Adversarial Examples Detection and Analysis with Layer-wise Autoencoders

We present a mechanism for detecting adversarial examples based on data representations taken from the hidden layers of the target network. For this purpose, we train individual autoencoders at intermediate layers of the target network. This allows us to describe the manifold of true data and, in consequence, decide whether a given example has the same characteristics as true data. It also gives us insight into the behavior of adversarial examples and their flow through the layers of a deep neural network. Experimental results show that our method outperforms the state of the art in supervised and unsupervised settings.

cs.LG↗

arXiv2020

HyperFlow: Representing 3D Objects as Surfaces

In this work, we present HyperFlow - a novel generative model that leverages hypernetworks to create continuous 3D object representations in a form of lightweight surfaces (meshes), directly out of point clouds. Efficient object representations are essential for many computer vision applications, including robotic manipulation and autonomous driving. However, creating those representations is often cumbersome, because it requires processing unordered sets of point clouds. Therefore, it is either computationally expensive, due to additional optimization constraints such as permutation invariance, or leads to quantization losses introduced by binning point clouds into discrete voxels. Inspired by mesh-based representations of objects used in computer graphics, we postulate a fundamentally different approach and represent 3D objects as a family of surfaces. To that end, we devise a generative model that uses a hypernetwork to return the weights of a Continuous Normalizing Flows (CNF) target network. The goal of this target network is to map points from a probability distribution into a 3D mesh. To avoid numerical instability of the CNF on compact support distributions, we propose a new Spherical Log-Normal function which models density of 3D points around object surfaces mimicking noise introduced by 3D capturing devices. As a result, we obtain continuous mesh-based object representations that yield better qualitative results than competing approaches, while reducing training time by over an order of magnitude.

cs.CV↗

arXiv2020

Finding the Optimal Network Depth in Classification Tasks

We develop a fast end-to-end method for training lightweight neural networks using multiple classifier heads. By allowing the model to determine the importance of each head and rewarding the choice of a single shallow classifier, we are able to detect and remove unneeded components of the network. This operation, which can be seen as finding the optimal depth of the model, significantly reduces the number of parameters and accelerates inference across different hardware processing units, which is not the case for many standard pruning methods. We show the performance of our method on multiple network architectures and datasets, analyze its optimization properties, and conduct ablation studies.

cs.LG↗

arXiv2020

The Break-Even Point on Optimization Trajectories of Deep Neural Networks

The early phase of training of deep neural networks is critical for their final performance. In this work, we study how the hyperparameters of stochastic gradient descent (SGD) used in the early phase of training affect the rest of the optimization trajectory. We argue for the existence of the "break-even" point on this trajectory, beyond which the curvature of the loss surface and noise in the gradient are implicitly regularized by SGD. In particular, we demonstrate on multiple classification tasks that using a large learning rate in the initial phase of training reduces the variance of the gradient, and improves the conditioning of the covariance of gradients. These effects are beneficial from the optimization perspective and become visible after the break-even point. Complementing prior work, we also show that using a low learning rate results in bad conditioning of the loss surface even for a neural network with batch normalization layers. In short, our work shows that key properties of the loss surface are strongly influenced by SGD in the early phase of training. We argue that studying the impact of the identified effects on generalization is a promising future direction.

cs.LG↗

arXiv2020

Molecule Attention Transformer

Designing a single neural network architecture that performs competitively across a range of molecule property prediction tasks remains largely an open challenge, and its solution may unlock a widespread use of deep learning in the drug discovery industry. To move towards this goal, we propose Molecule Attention Transformer (MAT). Our key innovation is to augment the attention mechanism in Transformer using inter-atomic distances and the molecular graph structure. Experiments show that MAT performs competitively on a diverse set of molecular prediction tasks. Most importantly, with a simple self-supervised pretraining, MAT requires tuning of only a few hyperparameter values to achieve state-of-the-art performance on downstream tasks. Finally, we show that attention weights learned by MAT are interpretable from the chemical point of view.

cs.LG↗

arXiv2019

Set Aggregation Network as a Trainable Pooling Layer

Global pooling, such as max- or sum-pooling, is one of the key ingredients in deep neural networks used for processing images, texts, graphs and other types of structured data. Based on the recent DeepSets architecture proposed by Zaheer et al. (NIPS 2017), we introduce a Set Aggregation Network (SAN) as an alternative global pooling layer. In contrast to typical pooling operators, SAN allows to embed a given set of features to a vector representation of arbitrary size. We show that by adjusting the size of embedding, SAN is capable of preserving the whole information from the input. In experiments, we demonstrate that replacing global pooling layer by SAN leads to the improvement of classification accuracy. Moreover, it is less prone to overfitting and can be used as a regularizer.

cs.LG↗

arXiv2019

Biologically-Inspired Spatial Neural Networks

We introduce bio-inspired artificial neural networks consisting of neurons that are additionally characterized by spatial positions. To simulate properties of biological systems we add the costs penalizing long connections and the proximity of neurons in a two-dimensional space. Our experiments show that in the case where the network performs two different tasks, the neurons naturally split into clusters, where each cluster is responsible for processing a different task. This behavior not only corresponds to the biological systems, but also allows for further insight into interpretability or continual learning.

cs.NE↗

arXiv2019

Fast and Stable Interval Bounds Propagation for Training Verifiably Robust Models

We present an efficient technique, which allows to train classification networks which are verifiably robust against norm-bounded adversarial attacks. This framework is built upon the work of Gowal et al., who applies the interval arithmetic to bound the activations at each layer and keeps the prediction invariant to the input perturbation. While that method is faster than competitive approaches, it requires careful tuning of hyper-parameters and a large number of epochs to converge. To speed up and stabilize training, we supply the cost function with an additional term, which encourages the model to keep the interval bounds at hidden layers small. Experimental results demonstrate that we can achieve comparable (or even better) results using a smaller number of training iterations, in a more stable fashion. Moreover, the proposed model is not so sensitive to the exact specification of the training process, which makes it easier to use by practitioners.

cs.LG↗

arXiv2019

Cramer-Wold AutoEncoder

We propose a new generative model, Cramer-Wold Autoencoder (CWAE). Following WAE, we directly encourage normality of the latent space. Our paper uses also the recent idea from Sliced WAE (SWAE) model, which uses one-dimensional projections as a method of verifying closeness of two distributions. The crucial new ingredient is the introduction of a new (Cramer-Wold) metric in the space of densities, which replaces the Wasserstein metric used in SWAE. We show that the Cramer-Wold metric between Gaussian mixtures is given by a simple analytic formula, which results in the removal of sampling necessary to estimate the cost function in WAE and SWAE models. As a consequence, while drastically simplifying the optimization procedure, CWAE produces samples of a matching perceptual quality to other SOTA models.

cs.LG↗

arXiv2019

Hypernetwork functional image representation

Motivated by the human way of memorizing images we introduce their functional representation, where an image is represented by a neural network. For this purpose, we construct a hypernetwork which takes an image and returns weights to the target network, which maps point from the plane (representing positions of the pixel) into its corresponding color in the image. Since the obtained representation is continuous, one can easily inspect the image at various resolutions and perform on it arbitrary continuous operations. Moreover, by inspecting interpolations we show that such representation has some properties characteristic to generative models. To evaluate the proposed mechanism experimentally, we apply it to image super-resolution problem. Despite using a single model for various scaling factors, we obtained results comparable to existing super-resolution methods.

cs.LG↗

arXiv2019

One-element Batch Training by Moving Window

Several deep models, esp. the generative, compare the samples from two distributions (e.g. WAE like AutoEncoder models, set-processing deep networks, etc) in their cost functions. Using all these methods one cannot train the model directly taking small size (in extreme -- one element) batches, due to the fact that samples are to be compared. We propose a generic approach to training such models using one-element mini-batches. The idea is based on splitting the batch in latent into parts: previous, i.e. historical, elements used for latent space distribution matching and the current ones, used both for latent distribution computation and the minimization process. Due to the smaller memory requirements, this allows to train networks on higher resolution images then in the classical approach.

cs.LG↗

arXiv2019

Independent Component Analysis based on multiple data-weighting

Independent Component Analysis (ICA) - one of the basic tools in data analysis - aims to find a coordinate system in which the components of the data are independent. In this paper we present Multiple-weighted Independent Component Analysis (MWeICA) algorithm, a new ICA method which is based on approximate diagonalization of weighted covariance matrices. Our idea is based on theoretical result, which says that linear independence of weighted data (for gaussian weights) guarantees independence. Experiments show that MWeICA achieves better results to most state-of-the-art ICA methods, with similar computational time.

cs.LG↗

arXiv2019

Processing of missing data by neural networks

We propose a general, theoretically justified mechanism for processing missing data by neural networks. Our idea is to replace typical neuron's response in the first hidden layer by its expected value. This approach can be applied for various types of networks at minimal cost in their modification. Moreover, in contrast to recent approaches, it does not require complete data for training. Experimental results performed on different types of architectures show that our method gives better results than typical imputation strategies and other methods dedicated for incomplete data.

cs.LG↗

arXiv2019

Dynamical Isometry is Achieved in Residual Networks in a Universal Way for any Activation Function

We demonstrate that in residual neural networks (ResNets) dynamical isometry is achievable irrespectively of the activation function used. We do that by deriving, with the help of Free Probability and Random Matrix Theories, a universal formula for the spectral density of the input-output Jacobian at initialization, in the large network width and depth limit. The resulting singular value spectrum depends on a single parameter, which we calculate for a variety of popular activation functions, by analyzing the signal propagation in the artificial neural network. We corroborate our results with numerical simulations of both random matrices and ResNets applied to the CIFAR-10 classification problem. Moreover, we study the consequence of this universal behavior for the initial and late phases of the learning processes. We conclude by drawing attention to the simple fact, that initialization acts as a confounding factor between the choice of activation function and the rate of learning. We propose that in ResNets this can be resolved based on our results, by ensuring the same level of dynamical isometry at initialization.

stat.ML↗

arXiv2019

Non-linear ICA based on Cramer-Wold metric

Non-linear source separation is a challenging open problem with many applications. We extend a recently proposed Adversarial Non-linear ICA (ANICA) model, and introduce Cramer-Wold ICA (CW-ICA). In contrast to ANICA we use a simple, closed--form optimization target instead of a discriminator--based independence measure. Our results show that CW-ICA achieves comparable results to ANICA, while foregoing the need for adversarial training.

cs.LG↗

arXiv2019

LOSSGRAD: automatic learning rate in gradient descent

In this paper, we propose a simple, fast and easy to implement algorithm LOSSGRAD (locally optimal step-size in gradient descent), which automatically modifies the step-size in gradient descent during neural networks training. Given a function $f$, a point $x$, and the gradient $\nabla_x f$ of $f$, we aim to find the step-size $h$ which is (locally) optimal, i.e. satisfies: $$ h=arg\,min_{t \geq 0} f(x-t \nabla_x f). $$ Making use of quadratic approximation, we show that the algorithm satisfies the above assumption. We experimentally show that our method is insensitive to the choice of initial learning rate while achieving results comparable to other methods.

cs.LG↗

arXiv2019

Sliced generative models

In this paper we discuss a class of AutoEncoder based generative models based on one dimensional sliced approach. The idea is based on the reduction of the discrimination between samples to one-dimensional case. Our experiments show that methods can be divided into two groups. First consists of methods which are a modification of standard normality tests, while the second is based on classical distances between samples. It turns out that both groups are correct generative models, but the second one gives a slightly faster decrease rate of Fréchet Inception Distance (FID).

cs.LG↗

arXiv2018

Cascade context encoder for improved inpainting

In this paper, we analyze if cascade usage of the context encoder with increasing input can improve the results of the inpainting. For this purpose, we train context encoder for 64x64 pixels images in a standard way and use its resized output to fill in the missing input region of the 128x128 context encoder, both in training and evaluation phase. As the result, the inpainting is visibly more plausible. In order to thoroughly verify the results, we introduce normalized squared-distortion, a measure for quantitative inpainting evaluation, and we provide its mathematical explanation. This is the first attempt to formalize the inpainting measure, which is based on the properties of latent feature representation, instead of L2 reconstruction loss.

cs.CV↗