We use distributionally-robust optimization for machine learning to mitigate the effect of data poisoning attacks. We provide performance guarantees for the trained model on the original data (not including the poison records) by training the model for the worst-case distribution on a neighbourhood around the empirical distribution (extracted from the training dataset corrupted by a poisoning attack) defined using the Wasserstein distance. We relax the distributionally-robust machine learning problem by finding an upper bound for the worst-case fitness based on the empirical sampled-averaged fitness and the Lipschitz-constant of the fitness function (on the data for given model parameters) as regularizer. For regression models, we prove that this regularizer is equal to the dual norm of the model parameters. We use the Wine Quality dataset, the Boston Housing Market dataset, and the Adult dataset for demonstrating the results of this paper.

Positive-confidence (Pconf) classification [Ishida et al., 2018] is a promising weakly-supervised learning method which trains a binary classifier only from positive data equipped with confidence. However, in practice, the confidence may be skewed by bias arising in an annotation process. The Pconf classifier cannot be properly learned with skewed confidence, and consequently, the classification performance might be deteriorated. In this paper, we introduce the parameterized model of the skewed confidence, and propose the method for selecting the hyperparameter which cancels out the negative impact of skewed confidence under the assumption that we have the misclassification rate of positive samples as a prior knowledge. We demonstrate the effectiveness of the proposed method through a synthetic experiment with simple linear models and benchmark problems with neural network models. We also apply our method to drivers' drowsiness prediction to show that it works well with a real-world problem where confidence is obtained based on manual annotation.

The use of drug combinations often leads to polypharmacy side effects (POSE). A recent method formulates POSE prediction as a link prediction problem on a graph of drugs and proteins, and solves it with Graph Convolutional Networks (GCNs). However, due to the complex relationships in POSE, this method has high computational cost and memory demand. This paper proposes a flexible Tri-graph Information Propagation (TIP) model that operates on three subgraphs to learn representations progressively by propagation from protein-protein graph to drug-drug graph via protein-drug graph. Experiments show that TIP improves accuracy by 7%+, time efficiency by 83$\times$, and space efficiency by 3$\times$.

--Deep neural networks achieve state-of-the-art performance for a range of classification and inference tasks. However, the use of stochastic gradient descent combined with the noncon-vexity of the underlying optimization problems renders parameter learning susceptible to initialization. T o address this issue, a variety of methods that rely on random parameter initialization or knowledge distillation have been proposed in the past. In this paper, we propose FuseInit, a novel method to initialize shallower networks by fusing neighboring layers of deeper networks that are trained with random initialization. We develop theoretical results and efficient algorithms for mean-square error (MSE)- optimal fusion of neighboring dense-dense, convolutional-dense, and convolutional-convolutional layers. We show experiments for a range of classification and regression datasets, which suggest that deeper neural networks are less sensitive to initialization and shallower networks can perform better (sometimes as well as their deeper counterparts) if initialized with FuseInit.

A wide class of quantum algorithms for learning problems exp loit fast quantum linear algebra subroutines to achieve runtimes that are exponentially faster than their classical counterparts [ Cil 18 ]. Examples of these algorithms are quantum support vector m achines [ RML14 ], quantum linear regression [ WBL12; SSP16 ], and quantum least squares [ KP17; CGJ18 ]. A careful analysis of these algorithms identified a number of caveats that limit their practical applicability such as the need for a strong form of quantum ac cess to the input data, restrictions on structural properties of the data matrix (such as conditi on number or sparsity), and modes of access to the output [ Aar15 ]. Furthermore, if one assumes that it is efficient to (classic ally) sample elements of the training data in a way proportional to their norm, then it is possible to show that classical algorithms are only polynomially slowe r (albeit the scaling of the quantum algorithms can be considerably better) [ Tan18; CL W18; Chi 19a; GLT18; Chi 19b ]. In this work we continue to investigate the limitations of qu antum algorithms for learning problems.

Machine learning techniques have been paramount throughout the last years, being applied in a wide range of tasks, such as classification, object recognition, person identification, image segmentation, among others. Nevertheless, conventional classification algorithms, e.g., Logistic Regression, Decision Trees, Bayesian classifiers, might lack complexity and diversity, not being suitable when dealing with real-world data. A recent graph-inspired classifier, known as the Optimum-Path Forest, has proven to be a state-of-the-art technique, comparable to Support Vector Machines and even surpassing it in some tasks. In this paper, we propose a Python-based Optimum-Path Forest framework, denoted as OPFython, where all of its functions and classes are based upon the original C language implementation. Additionally, as OPFython is a Python-based library, it provides a more friendly environment and a faster prototyping workspace than the C language.

Noisy labeled data is more a norm than a rarity for self-generated content that is continuously published on the web and social media. Due to privacy concerns and governmental regulations, such a data stream can only be stored and used for learning purposes in a limited duration. To overcome the noise in this on-line scenario we propose QActor which novel combines: the selection of supposedly clean samples via quality models and actively querying an oracle for the most informative true labels. While the former can suffer from low data volumes of on-line scenarios, the latter is constrained by the availability and costs of human experts. QActor swiftly combines the merits of quality models for data filtering and oracle queries for cleaning the most informative data. The objective of QActor is to leverage the stringent oracle budget to robustly maximize the learning accuracy. QActor explores various strategies combining different query allocations and uncertainty measures. A central feature of QActor is to dynamically adjust the query limit according to the learning loss for each data batch. We extensively evaluate different image datasets fed into the classifier that can be standard machine learning (ML) models or deep neural networks (DNN) with noise label ratios ranging between 30% and 80%. Our results show that QActor can nearly match the optimal accuracy achieved using only clean data at the cost of at most an additional 6% of ground truth data from the oracle.

We introduce a random forest approach to enable spreads' prediction in the primary catastrophe bond market. We investigate whether all information provided to investors in the offering circular prior to a new issuance is equally important in predicting its spread. The whole population of non-life catastrophe bonds issued from December 2009 to May 2018 is used. The random forest shows an impressive predictive power on unseen primary catastrophe bond data explaining 93% of the total variability. For comparison, linear regression, our benchmark model, has inferior predictive performance explaining only 47% of the total variability. All details provided in the offering circular are predictive of spread but in a varying degree. The stability of the results is studied. The usage of random forest can speed up investment decisions in the catastrophe bond industry.

As is typical in other fields of application of high throughput systems, radiology is faced with the challenge of interpreting increasingly sophisticated predictive models such as those derived from radiomics analyses. Interpretation may be guided by the learning output from machine learning models, which may however vary greatly with each technique. Whatever this output model, it will raise some essential questions. How do we interpret the prognostic model for clinical implementation? How can we identify potential information structures within sets of radiomic features, in order to create clinically interpretable models? And how can we recombine or exploit potential relationships between features towards improved interpretability? A number of statistical techniques are explored to assess (possibly nonlinear) relationships between radiological features from different angles.

APPENDIX A. Visualizing the Wishart Distribution The Wishart distribution is a generalization to multiple dimensions of the chi-squared distribution, or in the case of non-integer degrees of freedom, of the gamma distribution. We show in fig.5 that for a 1-dimensional and equal to 1 Σ scale matrix, the Wishart distribution W 1( n, 1) is equivalent to the χ 2 ( n) distribution. Figure 5: Monodimensional Wishart Distribution and χ 2 (n) distribution comparison Save for this simple case, being the Wishart a distribution over matrices, it is a generally hard task to visualize it as a density function. Samples can be however drawn from it and the eigenvectors and eigenvalues of the resulting sampled matrix can be exploited to define an ellipse. An example of this technique is shown in fig.6.