arXiv subjects

Pietro Lio

Publications and source records attributed to Pietro Lio.

At least 91 records · Page 5Linked to original sources

arXiv2021

GRADE: Graph Dynamic Embedding

Representation learning of static and more recently dynamically evolving graphs has gained noticeable attention. Existing approaches for modelling graph dynamics focus extensively on the evolution of individual nodes independently of the evolution of mesoscale community structures. As a result, current methods do not provide useful tools to study and cannot explicitly capture temporal community dynamics. To address this challenge, we propose GRADE - a probabilistic model that learns to generate evolving node and community representations by imposing a random walk prior over their trajectories. Our model also learns node community membership which is updated between time steps via a transition matrix. At each time step link generation is performed by first assigning node membership from a distribution over the communities, and then sampling a neighbor from a distribution over the nodes for the assigned community. We parametrize the node and community distributions with neural networks and learn their parameters via variational inference. Experiments demonstrate GRADE outperforms baselines in dynamic link prediction, shows favourable performance on dynamic community detection, and identifies coherent and interpretable evolving communities.

cs.LG↗

arXiv2020

Extrapolatable Relational Reasoning With Comparators in Low-Dimensional Manifolds

While modern deep neural architectures generalise well when test data is sampled from the same distribution as training data, they fail badly for cases when the test data distribution differs from the training distribution even along a few dimensions. This lack of out-of-distribution generalisation is increasingly manifested when the tasks become more abstract and complex, such as in relational reasoning. In this paper we propose a neuroscience-inspired inductive-biased module that can be readily amalgamated with current neural network architectures to improve out-of-distribution (o.o.d) generalisation performance on relational reasoning tasks. This module learns to project high-dimensional object representations to low-dimensional manifolds for more efficient and generalisable relational comparisons. We show that neural nets with this inductive bias achieve considerably better o.o.d generalisation performance for a range of relational reasoning tasks. We finally analyse the proposed inductive bias module to understand the importance of lower dimension projection, and propose an augmentation to the algorithmic alignment theory to better measure algorithmic alignment with generalisation.

cs.LG↗

arXiv2020

Learned Low Precision Graph Neural Networks

Deep Graph Neural Networks (GNNs) show promising performance on a range of graph tasks, yet at present are costly to run and lack many of the optimisations applied to DNNs. We show, for the first time, how to systematically quantise GNNs with minimal or no loss in performance using Network Architecture Search (NAS). We define the possible quantisation search space of GNNs. The proposed novel NAS mechanism, named Low Precision Graph NAS (LPGNAS), constrains both architecture and quantisation choices to be differentiable. LPGNAS learns the optimal architecture coupled with the best quantisation strategy for different components in the GNN automatically using back-propagation in a single search round. On eight different datasets, solving the task of classifying unseen nodes in a graph, LPGNAS generates quantised models with significant reductions in both model and buffer sizes but with similar accuracy to manually designed networks and other NAS results. In particular, on the Pubmed dataset, LPGNAS shows a better size-accuracy Pareto frontier compared to seven other manual and searched baselines, offering a 2.3 times reduction in model size but a 0.4% increase in accuracy when compared to the best NAS competitor. Finally, from our collected quantisation statistics on a wide range of datasets, we suggest a W4A8 (4-bit weights, 8-bit activations) quantisation strategy might be the bottleneck for naive GNN quantisations.

cs.LG↗

arXiv2020

Learning distributed representations of graphs with Geo2DR

We present Geo2DR (Geometric to Distributed Representations), a GPU ready Python library for unsupervised learning on graph-structured data using discrete substructure patterns and neural language models. It contains efficient implementations of popular graph decomposition algorithms and neural language models in PyTorch which can be combined to learn representations of graphs using the distributive hypothesis. Furthermore, Geo2DR comes with general data processing and loading methods to bring substantial speed-up in the training of the neural language models. Through this we provide a modular set of tools and methods to quickly construct systems capable of learning distributed representations of graphs. This is useful for replication of existing methods, modification, or development of completely new methods. This paper serves to present the Geo2DR library and perform a comprehensive comparative analysis of existing methods re-implemented using Geo2DR across widely used graph classification benchmarks. Geo2DR displays a high reproducibility of results in published methods and interoperability with other libraries useful for distributive language modelling.

cs.LG↗

arXiv2020

Path Integral Based Convolution and Pooling for Graph Neural Networks

Graph neural networks (GNNs) extends the functionality of traditional neural networks to graph-structured data. Similar to CNNs, an optimized design of graph convolution and pooling is key to success. Borrowing ideas from physics, we propose a path integral based graph neural networks (PAN) for classification and regression tasks on graphs. Specifically, we consider a convolution operation that involves every path linking the message sender and receiver with learnable weights depending on the path length, which corresponds to the maximal entropy random walk. It generalizes the graph Laplacian to a new transition matrix we call maximal entropy transition (MET) matrix derived from a path integral formalism. Importantly, the diagonal entries of the MET matrix are directly related to the subgraph centrality, thus providing a natural and adaptive pooling mechanism. PAN provides a versatile framework that can be tailored for different graph data with varying sizes and structures. We can view most existing GNN architectures as special cases of PAN. Experimental results show that PAN achieves state-of-the-art performance on various graph classification/regression tasks, including a new benchmark dataset from statistical mechanics we propose to boost applications of GNN in physical sciences.

cs.LG↗

arXiv2020

Abstract Diagrammatic Reasoning with Multiplex Graph Networks

Abstract reasoning, particularly in the visual domain, is a complex human ability, but it remains a challenging problem for artificial neural learning systems. In this work we propose MXGNet, a multilayer graph neural network for multi-panel diagrammatic reasoning tasks. MXGNet combines three powerful concepts, namely, object-level representation, graph neural networks and multiplex graphs, for solving visual reasoning tasks. MXGNet first extracts object-level representations for each element in all panels of the diagrams, and then forms a multi-layer multiplex graph capturing multiple relations between objects across different diagram panels. MXGNet summarises the multiple graphs extracted from the diagrams of the task, and uses this summarisation to pick the most probable answer from the given candidates. We have tested MXGNet on two types of diagrammatic reasoning tasks, namely Diagram Syllogisms and Raven Progressive Matrices (RPM). For an Euler Diagram Syllogism task MXGNet achieves state-of-the-art accuracy of 99.8%. For PGM and RAVEN, two comprehensive datasets for RPM reasoning, MXGNet outperforms the state-of-the-art models by a considerable margin.

cs.LG↗

arXiv2020

ML-SIM: A deep neural network for reconstruction of structured illumination microscopy images

Structured illumination microscopy (SIM) has become an important technique for optical super-resolution imaging because it allows a doubling of image resolution at speeds compatible for live-cell imaging. However, the reconstruction of SIM images is often slow and prone to artefacts. Here we propose a versatile reconstruction method, ML-SIM, which makes use of machine learning. The model is an end-to-end deep residual neural network that is trained on a simulated data set to be free of common SIM artefacts. ML-SIM is thus robust to noise and irregularities in the illumination patterns of the raw SIM input frames. The reconstruction method is widely applicable and does not require the acquisition of experimental training data. Since the training data are generated from simulations of the SIM process on images from generic libraries the method can be efficiently adapted to specific experimental SIM implementations. The reconstruction quality enabled by our method is compared with traditional SIM reconstruction methods, and we demonstrate advantages in terms of noise, reconstruction fidelity and contrast for both simulated and experimental inputs. In addition, reconstruction of one SIM frame typically only takes ~100ms to perform on PCs with modern Nvidia graphics cards, making the technique compatible with real-time imaging. The full implementation and the trained networks are available at http://ML-SIM.com.

eess.IV↗

arXiv2020

Probabilistic Dual Network Architecture Search on Graphs

We present the first differentiable Network Architecture Search (NAS) for Graph Neural Networks (GNNs). GNNs show promising performance on a wide range of tasks, but require a large amount of architecture engineering. First, graphs are inherently a non-Euclidean and sophisticated data structure, leading to poor adaptivity of GNN architectures across different datasets. Second, a typical graph block contains numerous different components, such as aggregation and attention, generating a large combinatorial search space. To counter these problems, we propose a Probabilistic Dual Network Architecture Search (PDNAS) framework for GNNs. PDNAS not only optimises the operations within a single graph block (micro-architecture), but also considers how these blocks should be connected to each other (macro-architecture). The dual architecture (micro- and marco-architectures) optimisation allows PDNAS to find deeper GNNs on diverse datasets with better performance compared to other graph NAS methods. Moreover, we use a fully gradient-based search approach to update architectural parameters, making it the first differentiable graph NAS method. PDNAS outperforms existing hand-designed GNNs and NAS results, for example, on the PPI dataset, PDNAS beats its best competitors by 1.67 and 0.17 in F1 scores.

cs.LG↗

arXiv2019

Decoupling feature propagation from the design of graph auto-encoders

We present two instances, L-GAE and L-VGAE, of the variational graph auto-encoding family (VGAE) based on separating feature propagation operations from graph convolution layers typically found in graph learning methods to a single linear matrix computation made prior to input in standard auto-encoder architectures. This decoupling enables the independent and fixed design of the auto-encoder without requiring additional GCN layers for every desired increase in the size of a node's local receptive field. Fixing the auto-encoder enables a fairer assessment on the size of a nodes receptive field in building representations. Furthermore a by-product of fixing the auto-encoder design often results in substantially smaller networks than their VGAE counterparts especially as we increase the number of feature propagations. A comparative downstream evaluation on link prediction tasks show comparable state of the art performance to similar VGAE arrangements despite considerable simplification. We also show the simple application of our methodology to more challenging representation learning scenarios such as spatio-temporal graph representation learning.

cs.LG↗

arXiv2019

Dynamic Neural Network Channel Execution for Efficient Training

Existing methods for reducing the computational burden of neural networks at run-time, such as parameter pruning or dynamic computational path selection, focus solely on improving computational efficiency during inference. On the other hand, in this work, we propose a novel method which reduces the memory footprint and number of computing operations required for training and inference. Our framework efficiently integrates pruning as part of the training procedure by exploring and tracking the relative importance of convolutional channels. At each training step, we select only a subset of highly salient channels to execute according to the combinatorial upper confidence bound algorithm, and run a forward and backward pass only on these activated channels, hence learning their parameters. Consequently, we enable the efficient discovery of compact models. We validate our approach empirically on state-of-the-art CNNs - VGGNet, ResNet and DenseNet, and on several image classification datasets. Results demonstrate our framework for dynamic channel execution reduces computational cost up to 4x and parameter count up to 9x, thus reducing the memory and computational demands for discovering and training compact neural network models.

cs.LG↗

arXiv2019

Unsupervised and interpretable scene discovery with Discrete-Attend-Infer-Repeat

In this work we present Discrete Attend Infer Repeat (Discrete-AIR), a Recurrent Auto-Encoder with structured latent distributions containing discrete categorical distributions, continuous attribute distributions, and factorised spatial attention. While inspired by the original AIR model andretaining AIR model's capability in identifying objects in an image, Discrete-AIR provides direct interpretability of the latent codes. We show that for Multi-MNIST and a multiple-objects version of dSprites dataset, the Discrete-AIR model needs just one categorical latent variable, one attribute variable (for Multi-MNIST only), together with spatial attention variables, for efficient inference. We perform analysis to show that the learnt categorical distributions effectively capture the categories of objects in the scene for Multi-MNIST and for Multi-Sprites.

cs.CV↗

arXiv2019

How Can We Make GAN Perform Better in Single Medical Image Super-Resolution? A Lesion Focused Multi-Scale Approach

Single image super-resolution (SISR) is of great importance as a low-level computer vision task. The fast development of Generative Adversarial Network (GAN) based deep learning architectures realises an efficient and effective SISR to boost the spatial resolution of natural images captured by digital cameras. However, the SISR for medical images is still a very challenging problem. This is due to (1) compared to natural images, in general, medical images have lower signal to noise ratios, (2) GAN based models pre-trained on natural images may synthesise unrealistic patterns in medical images which could affect the clinical interpretation and diagnosis, and (3) the vanilla GAN architecture may suffer from unstable training and collapse mode that can also affect the SISR results. In this paper, we propose a novel lesion focused SR (LFSR) method, which incorporates GAN to achieve perceptually realistic SISR results for brain tumour MRI images. More importantly, we test and make comparison using recently developed GAN variations, e.g., Wasserstein GAN (WGAN) and WGAN with Gradient Penalty (WGAN-GP), and propose a novel multi-scale GAN (MS-GAN), to achieve a more stabilised and efficient training and improved perceptual quality of the super-resolved results. Based on both quantitative evaluations and our designed mean opinion score, the proposed LFSR coupled with MS-GAN has performed better in terms of both perceptual quality and efficiency.

eess.IV↗

arXiv2018

Unseen Word Representation by Aligning Heterogeneous Lexical Semantic Spaces

Word embedding techniques heavily rely on the abundance of training data for individual words. Given the Zipfian distribution of words in natural language texts, a large number of words do not usually appear frequently or at all in the training data. In this paper we put forward a technique that exploits the knowledge encoded in lexical resources, such as WordNet, to induce embeddings for unseen words. Our approach adapts graph embedding and cross-lingual vector space transformation techniques in order to merge lexical knowledge encoded in ontologies with that derived from corpus statistics. We show that the approach can provide consistent performance improvements across multiple evaluation benchmarks: in-vitro, on multiple rare word similarity datasets, and in-vivo, in two downstream text classification tasks.

cs.CL↗

arXiv2018

Lesion Focused Super-Resolution

Super-resolution (SR) for image enhancement has great importance in medical image applications. Broadly speaking, there are two types of SR, one requires multiple low resolution (LR) images from different views of the same object to be reconstructed to the high resolution (HR) output, and the other one relies on the learning from a large amount of training datasets, i.e., LR-HR pairs. In real clinical environment, acquiring images from multi-views is expensive and sometimes infeasible. In this paper, we present a novel Generative Adversarial Networks (GAN) based learning framework to achieve SR from its LR version. By performing simulation based studies on the Multimodal Brain Tumor Segmentation Challenge (BraTS) datasets, we demonstrate the efficacy of our method in application of brain tumor MRI enhancement. Compared to bilinear interpolation and other state-of-the-art SR methods, our model is lesion focused, which is not only resulted in better perceptual image quality without blurring, but also more efficient and directly benefit for the following clinical tasks, e.g., lesion detection and abnormality enhancement. Therefore, we can envisage the application of our SR method to boost image spatial resolution while maintaining crucial diagnostic information for further clinical tasks.

eess.IV↗

arXiv2018

Multi-omic Network Regression: Methodology, Tool and Case Study

The analysis of biological networks is characterized by the definition of precise linear constraints used to cumulatively reduce the solution space of the computed states of a multi-omic (for instance metabolic, transcriptomic and proteomic) model. In this paper, we attempt, for the first time, to combine metabolic modelling and networked Cox regression, using the metabolic model of the bacterium Helicobacter Pylori. This enables a platform both for quantitative analysis of networked regression, but also testing the findings from network regression (a list of significant vectors and their networked relationships) on in vivo transcriptomic data. Data generated from the model, using flux balance analysis to construct a Pareto front, specifically, a trade-off of Oxygen exchange and growth rate and a trade-off of Carbon Dioxide exchange and growth rate, is analysed and then the model is used to quantify the success of the analysis. It was found that using the analysis, reconstruction of the initial data was considerably more successful than a pure noise alternative. Our methodological approach is quite general and it could be of interest for the wider community of complex networks researchers; it is implemented in a software tool, MoNeRe, which is freely available through the Github platform.

q-bio.MN↗

arXiv2010

Noise and nonlinearities in high-throughput data

High-throughput data analyses are becoming common in biology, communications, economics and sociology. The vast amounts of data are usually represented in the form of matrices and can be considered as knowledge networks. Spectra-based approaches have proved useful in extracting hidden information within such networks and for estimating missing data, but these methods are based essentially on linear assumptions. The physical models of matching, when applicable, often suggest non-linear mechanisms, that may sometimes be identified as noise. The use of non-linear models in data analysis, however, may require the introduction of many parameters, which lowers the statistical weight of the model. According to the quality of data, a simpler linear analysis may be more convenient than more complex approaches. In this paper, we show how a simple non-parametric Bayesian model may be used to explore the role of non-linearities and noise in synthetic and experimental data sets.

q-bio.QM↗

arXiv2007

Risk perception in epidemic modeling

We investigate the effects of risk perception in a simple model of epidemic spreading. We assume that the perception of the risk of being infected depends on the fraction of neighbors that are ill. The effect of this factor is to decrease the infectivity, that therefore becomes a dynamical component of the model. We study the problem in the mean-field approximation and by numerical simulations for regular, random and scale-free networks. We show that for homogeneous and random networks, there is always a value of perception that stops the epidemics. In the ``worst-case'' scenario of a scale-free network with diverging input connectivity, a linear perception cannot stop the epidemics; however we show that a non-linear increase of the perception risk may lead to the extinction of the disease. This transition is discontinuous, and is not predicted by the mean-field analysis.

q-bio.PE↗

arXiv2006

Mathematical Model of HIV superinfection dynamics and R5 to X4 switch

During the HIV infection several quasispecies of the virus arise, which are able to use different coreceptors, in particular the CCR5 and CXCR4 coreceptors (R5 and X4 phenotypes, respectively). The switch in coreceptor usage has been correlated with a faster progression of the disease to the AIDS phase. As several pharmaceutical companies are starting large phase III trials for R5 and X4 drugs, models are needed to predict the co-evolutionary and competitive dynamics of virus strains. We present a model of HIV early infection which describes the dynamics of R5 quasispecies and a model of HIV late infection which describes the R5 to X4 switch. We report the following findings: after superinfection or coinfection, quasispecies dynamics has time scales of several months and becomes even slower at low number of CD4+ T cells. The curve of CD4+ T cells decreases, during AIDS late stage, and can be described taking into account the X4 related Tumor Necrosis Factor dynamics. Phylogenetic inference of chemokine receptors suggests that viral mutational pathway may generate R5 variants able to interact with chemokine receptors different from CXCR4. This may explain the massive signaling disruptions in the immune system observed during AIDS late stages and may have relevance for vaccination and therapy.

q-bio.PE↗