Advances in Neural Information Processing Systems 29 (NIPS 2016)
We propose a novel approach to reduce the computational cost of evaluation of convolutional neural networks, a factor that has hindered their deployment in low-power devices such as mobile phones. Inspired by the loop perforation technique from source code optimization, we speed up the bottleneck convolutional layers by skipping their evaluation in some of the spatial positions. We propose and analyze several strategies of choosing these positions. We demonstrate that perforation can accelerate modern convolutional networks such as AlexNet and VGG-16 by a factor of 2x - 4x. Additionally, we show that perforation is complementary to the recently proposed acceleration method of Zhang et al.
We consider a task of predicting normal and pathological phenotypes from macroscale human brain networks. These networks (connectomes) represent aggregated neural pathways between brain regions. We point to properties of connectomes that make them different from graphs arising in other application areas of network science. We discuss how machine learning can be organized on brain networks and focus on kernel classification methods. We describe different kernels on brain networks, including those that use information about similarity in spectral distributions of brain graphs and distances between optimal partitions of connectomes. We compare performance of the reviewed kernels in tasks of classifying autism spectrum disorder versus typical development and carriers versus non-carriers of an allele associated with an increased risk of Alzheimer’s disease.
CVPR is the premier annual computer vision event comprising the main conference and several co-located workshops and short courses. With its high quality and low cost, it provides an exceptional value for students, academics and industry researchers
We propose the model of Judge AI implemented by the approach based on commutative diagram of mappings. We suggest using the constellation for the grid of normative acts and the narrative together with the corresponding divergence. In this paper, the pre-trained models known in NLP as doc2vec and Fast Text are described. The evaluation of the quality of models car-ried out using open databases of court decisions.
Reservoir Computing (RC) is taking attention of neural networks structures developers because of machine learning algorithms are simple at the high level of generalization of the models. The approaches are numerous. RC can be applied to different architectures including recurrent neural networks with irregular connections that are called Echo State Networks (ESN). However, the existence of successful examples of chaotic sequences predictions does not provide successful method of multiple attribute objects classification.
In this paper the binary ESN classifiers are researched. We show that the reason of low precision of classification is the existence of unbalanced classes. Then the method to solve the problem is proposed. It is possible to use randomizing algorithm of learning data set balancing and method of data temporalization. The resulting errors matrixes have pretty good numbers. The proposed method is illustrated by the usage on synthetic data set. The features of ESN classifier are demonstrated in the case of rare events detection such as transaction attributes fraud detection.
This valuable source for graduate students and researchers provides a comprehensive introduction to current theories and applications in optimization methods and network models. Contributions to this book are focused on new efficient algorithms and rigorous mathematical theories, which can be used to optimize and analyze mathematical graph structures with massive size and high density induced by natural or artificial complex networks. Applications to social networks, power transmission grids, telecommunication networks, stock market networks, and human brain networks are presented.
Chapters in this book cover the following topics:Linear max min fairness Heuristic approaches for high-quality solutions Efficient approaches for complex multi-criteria optimization problems Comparison of heuristic algorithms New heuristic iterative local search Power in network structures Clustering nodes in random graphs Power transmission grid structure Network decomposition problems Homogeneity hypothesis testing Network analysis of international migration Social networks with node attributes Testing hypothesis on degree distribution in the market graphs Machine learning applications to human brain network studies
This proceeding is a result of The 6th International Conference on Network Analysis held at the Higher School of Economics, Nizhny Novgorod in May 2016. The conference brought together scientists and engineers from industry, government, and academia to discuss the links between network analysis and a variety of fields.
Interpretability and fairness are critical in computer vision and machine learning applications, in particular when dealing with human outcomes, e.g. inviting or not inviting for a job interview based on application materials that may include photographs. One promising direction to achieve fairness is by learning data representations that remove the semantics of protected characteristics, and are therefore able to mitigate unfair outcomes. All available models however learn latent embeddings which comes at the cost of being uninterpretable. We propose to cast this problem as data-to-data translation, i.e. learning a mapping from an input domain to a fair target domain, where a fairness definition is being enforced. Here the data domain can be images, or any tabular data representation. This task would be straightforward if we had fair target data available, but this is not the case. To overcome this, we learn a highly unconstrained mapping by exploiting statistics of residuals – the difference between input data and its translated version – and the protected characteristics. When applied to the CelebA dataset of face images with gender attribute as the protected characteristic, our model enforces equality of opportunity by adjusting the eyes and lips regions. Intriguingly, on the same dataset we arrive at similar conclusions when using semantic attribute representations of images for translation. On face images of the recent DiF dataset, with the same gender attribute, our method adjusts nose regions. In the Adult income dataset, also with protected gender attribute, our model achieves equality of opportunity by, among others, obfuscating the wife and husband relationship. Analyzing those systematic changes will allow us to scrutinize the interplay of fairness criterion, chosen protected characteristics, and prediction performance.
his volume presents the results of the Neural Information Processing Systems Competition track at the 2018 NeurIPS conference. The competition follows the same format as the 2017 competition track for NIPS. Out of 21 submitted proposals, eight competition proposals were selected, spanning the area of Robotics, Health, Computer Vision, Natural Language Processing, Systems and Physics.
Competitions have become an integral part of advancing state-of-the-art in artificial intelligence (AI). They exhibit one important difference to benchmarks: Competitions test a system end-to-end rather than evaluating only a single component; they assess the practicability of an algorithmic solution in addition to assessing feasibility.
This paper proposes a semi-conditional normalizing flow model for semi-supervised learning. The model uses both labeled and unlabeled data to learn an explicit model of joint distribution over objects and labels. Semi-conditional architecture of the model allows us to efficiently compute a value and gradients of the marginal likelihood for unlabeled objects. The conditional part of the model is based on a proposed conditional coupling layer. We demonstrate performance of the model for semi-supervised classification problem on different datasets. The model outperforms the baseline approach based on variational auto-encoders on MNIST dataset.
This paper is devoted to comparison of the capabilities of various methods to predict the bankruptcy of construction industry companies on a one-year horizon. The authors considered the following algorithms: logit and probit models, classification trees, random forests, artificial neural networks. Special attention was paid to the peculiarities of the training machine learning models, the impact of data imbalance on the predictive ability of models, analysis of ways to deal with these imbalances and analysis of the influence of non-financial factors on the predictive ability of models. In their study, the authors used non-financial and financial indicators calculated on the basis of public financial statements of the construction companies for the period from 2011 to 2017. The authors concluded that the models considered show acceptable quality for use in forecasting bankruptcy problems. The Gini or AUC coefficient (area under the ROC curve) was used as the quality markers of the model. It was revealed that neural networks outperform other methods in predictive power, while logistic regression models in combination with discretization follow them closely. It was found that the effective way to deal with the imbalance data depends on the type of model used. However, no significant impact on the imbalance in the training set predictive ability of the model was identified. The significant impact of nonfinancial indicators on the likelihood of bankruptcy was not confirmed.