1,721,043 research outputs found
Cluster partitioning in image analysis classification: A genetic algorithm approach
No full textA classification of data by using the genetic algorithm computational paradigm is proposed. The best data partition is defined to be the one minimizing the sum of Pythagorean distances between each datum in a cluster and the relative center of class or center of mass. Background is given, and the relevant genetic algorithm description is provided. The model for the genetic application is presented. Simulation results confirm genetic algorithms to be powerful tools for the solution of optimization problems
Graph Neural Networks in TensorFlow and Keras with Spektral
Full text linkGraph neural networks have enabled the application of deep learning to problems that can be described by graphs, which are found throughout the different fields of science, from physics to biology, natural language processing, telecommunications or medicine. In this paper we present Spektral, an open-source Python library for building graph neural networks with TensorFlow and the Keras application programming interface. Spektral implements a large set of methods for deep learning on graphs, including message-passing and pooling operators, as well as utilities for processing graphs and loading popular benchmark datasets. The purpose of this library is to provide the essential building blocks for creating graph neural networks, focusing on the guiding principles of user-fr iendliness and quick prototyping on which Keras is based. Spektral is, therefore, suitable for absolute beginners and expert deep learning practitioners alike. In this work, we present an overview of Spektral’s features and report the performance of the methods implemented by the library in scenarios of node classification, graph classification, and graph regression
Change-point methods on a sequence of graphs
Full text linkThis is the author accepted manuscript. The final version is available from the publisher via the DOI in this recordGiven a finite sequence of graphs, e.g. coming from
technological, biological, and social networks, the paper proposes
a methodology to identify possible changes in stationarity in the
stochastic process that generated such graphs. We consider a
general family of attributed graphs for which both topology
(vertices and edges) and associated attributes are allowed to
change over time, without violating the stationarity hypothesis.
Novel Change-Point Methods (CPMs) are proposed that map
graphs onto vectors, apply a suitable statistical test in vector
space and detect changes –if any– according to a user-defined
confidence level; an estimate for the change point is provided
as well. In particular, we propose two multivariate CPMs: one
designed to detect shifts in the mean, the other to address
more complex changes affecting the distribution. We ground
our methods on theoretical results that show how the inference
in the numerical vector space is related to the one in graph
domain, and vice-versa. We also extend the methodology to
handle multiple changes occurring in a single sequence. Results
show the effectiveness of what proposed in relevant application
scenarios.Swiss National Science Foundatio
Adversarial autoencoders with constant-curvature latent manifolds
Full text linkThis is the author accepted manuscript. The final version is available from the publisher via the DOI in this recordConstant-curvature Riemannian manifolds (CCMs)have been shown to be ideal embedding spaces in many application domains, as their non-Euclidean geometry can naturally account for some relevant properties of data, like hierarchy and circularity. In this work, we introduce the CCM adversarial autoencoder (CCM-AAE), a probabilistic generative model trained to represent a data distribution on a CCM. Our method works by matching the aggregated posterior of the CCM-AAE with a probability distribution defined on a CCM, so that the encoder implicitly learns to represent data on the CCM to fool the discriminator network. The geometric constraint is also explicitly imposed by jointly training the CCM-AAE to maximise the membership degree of the embeddings to the CCM. While a few works in recent literature make use of either hyperspherical or hyperbolic manifolds for different learning tasks, ours is the first unified framework to seamlessly deal with CCMs of different curvatures. We show the effectiveness of our model on three different datasets characterised by non-trivial geometry: semi-supervised classification on MNIST, link prediction on two popular citation datasets, and graph-based molecule generation using the QM9 chemical database. Results show that our method improves upon other autoencoders based on Euclidean and non-Euclidean geometries on all tasks taken into account.Swiss National Science Foundatio
Echo State Networks with Self-Normalizing Activations on the Hyper-Sphere
Full text linkThis is the final version. Available on open access from Nature Research via the DOI in this recordAmong the various architectures of Recurrent Neural Networks, Echo State Networks (ESNs) emerged due to their simplified and inexpensive training procedure. These networks are known to be sensitive to the setting of hyper-parameters, which critically affect their behaviour. Results show that their performance is usually maximized in a narrow region of hyper-parameter space called edge of chaos. Finding such a region requires searching in hyper-parameter space in a sensible way: hyper-parameter configurations marginally outside such a region might yield networks exhibiting fully developed chaos, hence producing unreliable computations. The performance gain due to optimizing hyper-parameters can be studied by considering the memory--nonlinearity trade-off, i.e., the fact that increasing the nonlinear behavior of the network degrades its ability to remember past inputs, and vice-versa. In this paper, we propose a model of ESNs that eliminates critical dependence on hyper-parameters, resulting in networks that provably cannot enter a chaotic regime and, at the same time, denotes nonlinear behaviour in phase space characterised by a large memory of past inputs, comparable to the one of linear networks. Our contribution is supported by experiments corroborating our theoretical findings, showing that the proposed model displays dynamics that are rich-enough to approximate many common nonlinear systems used for benchmarking.Canada Research Chairs progra
Distributed Deep Convolutional Neural Networks for the Internet-of-Things
No full textSevere constraints on memory and computation characterizing the Internet-of-Things (IoT) units may prevent the execution of Deep Learning (DL)-based solutions, which typically demand large memory and high processing load. In order to support a real-time execution of the considered DL model at the IoT unit level, DL solutions must be designed having in mind constraints on memory and processing capability exposed by the chosen IoT technology. In this paper, we introduce a design methodology aiming at allocating the execution of Convolutional Neural Networks (CNNs) on a distributed IoT application. Such a methodology is formalized as an optimization problem where the latency between the data-gathering phase and the subsequent decision-making one is minimized, within the given constraints on memory and processing load at the units level. The methodology supports multiple sources of data as well as multiple CNNs in execution on the same IoT system allowing the design of CNN-based applications demanding autonomy, low decision-latency, and high Quality-of-Service
Graph iForest: Isolation of anomalous and outlier graphs
No full textWe present an anomaly and outlier detection method for graph data. The method relies on the consideration that anomalies and outliers are more easily isolated by certain incremental partitionings of the data space. Specifically, we build upon the isolation forest method and introduce a new incremental partitioning of the space of graphs that makes the isolation forest method applicable to generic attributed graphs, i.e., graphs where both nodes and edges can be associated with attributes. Within the considered general setup, the topology and the number of nodes can change from graph to graph, and a node correspondence between different graphs can be absent or unknown. Examples of applications of what proposed include the identification of frauds and fake news in communication networks, and breakage of systems monitored by sensor networks. The main novel contribution of the paper is a graph space partitioning which we prove to be expressive enough to identify anomalies and outlier graphs in a given dataset. An empirical analysis on synthetic and real-world graphs validates the effectiveness of the proposed method
Cluster-based Aggregate Load Forecasting with Deep Neural Networks
No full textHighly accurate power demand forecasting represents one of key challenges of Smart Grid applications. In this setting, a large number of Smart Meters produces huge amounts of data that need to be processed to predict the load requested by the grid. Due to the high dimensionality of the problem, this often results in the adoption of simple aggregation strategies for the power that fail in capturing the relational information existing among the different types of user. A possible alternative, known as Cluster-based Aggregate Forecasting, consists in clustering the load profiles and, on top of that, building predictors of the aggregate at the cluster-level. In this work we explore the technique in the context of predictors based on deep recurrent neural networks and address the scalability issues presenting neural architectures adequate to process cluster-level aggregates. The proposed methods are finally evaluated both on a publicly available benchmark and a heterogenous dataset of Smart Meter data from an entire, medium-sized, Swiss town
Autoregressive Models for Sequences of Graphs
Full text linkThis paper proposes an autoregressive (AR) model for sequences of graphs, which generalises traditional AR models. A first novelty consists in formalising the AR model for a very general family of graphs, characterised by a variable topology, and attributes associated with nodes and edges. A graph neural network (GNN) is also proposed to learn the AR function associated with the graph-generating process (GGP), and subsequently predict the next graph in a sequence. The proposed method is compared with four baselines on synthetic GGPs, denoting a significantly better performance on all considered problems
- …
