We introduce a self-supervised vision representation model BEiT, which stands for Bidirectional Encoder representation from Image Transformers. Following BERT developed in the natural language processing area, we propose a masked image modeling task to pretrain vision Transformers. Specifically, each image has two views in our pre-training, i.e, image patches (such as 16x16 pixels), and visual tokens (i.e., discrete tokens). We first "tokenize" the original image into visual tokens. Then we randomly mask some image patches and fed them into the backbone Transformer. The pre-training objective is to recover the original visual tokens based on the corrupted image patches. After pre-training BEiT, we directly fine-tune the model parameters on downstream tasks by appending task layers upon the pretrained encoder. Experimental results on image classification and semantic segmentation show that our model achieves competitive results with previous pre-training methods. For example, base-size BEiT achieves 83.2% top-1 accuracy on ImageNet-1K, significantly outperforming from-scratch DeiT training (81.8%) with the same setup. Moreover, large-size BEiT obtains 86.3% only using ImageNet-1K, even outperforming ViT-L with supervised pre-training on ImageNet-22K (85.2%). The code and pretrained models are available at https://aka.ms/beit.

Top-k error has become a popular metric for large-scale classification benchmarks due to the inevitable semantic ambiguity among classes. Existing literature on top-k optimization generally focuses on the optimization method of the top-k objective, while ignoring the limitations of the metric itself. In this paper, we point out that the top-k objective lacks enough discrimination such that the induced predictions may give a totally irrelevant label a top rank. To fix this issue, we develop a novel metric named partial Area Under the top-k Curve (AUTKC). Theoretical analysis shows that AUTKC has a better discrimination ability, and its Bayes optimal score function could give a correct top-K ranking with respect to the conditional probability. This shows that AUTKC does not allow irrelevant labels to appear in the top list. Furthermore, we present an empirical surrogate risk minimization framework to optimize the proposed metric. Theoretically, we present (1) a sufficient condition for Fisher consistency of the Bayes optimal score function; (2) a generalization upper bound which is insensitive to the number of classes under a simple hyperparameter setting. Finally, the experimental results on four benchmark datasets validate the effectiveness of our proposed framework.

This paper constitutes the novel hypergraph convolutional neural networkbased clustering technique. This technique is employed to solve the clustering problem for the Citeseer dataset and the Cora dataset. Each dataset contains the feature matrix and the incidence matrix of the hypergraph (i.e., constructed from the feature matrix). This novel clustering method utilizes both matrices. Initially, the hypergraph auto-encoders are employed to transform both the incidence matrix and the feature matrix from high dimensional space to low dimensional space. In the end, we apply the k-means clustering technique to the transformed matrix. The hypergraph convolutional neural network (CNN)-based clustering technique presented a better result on performance during experiments than those of the other classical clustering techniques.

A Recurrent Neural Network that operates on several time lags, called an RNN(p), is the natural generalization of an Autoregressive ARX(p) model. It is a powerful forecasting tool when different time scales can influence a given phenomenon, as it happens in the energy sector where hourly, daily, weekly and yearly interactions coexist. The cost-effective BPTT is the industry standard as learning algorithm for RNNs. We prove that, when training RNN(p) models, other learning algorithms turn out to be much more efficient in terms of both time and space complexity. We also introduce a new learning algorithm, the Tree Recombined Recurrent Learning, that leverages on a tree representation of the unrolled network and appears to be even more effective. We present an application of RNN(p) models for power consumption forecasting on the hourly scale: experimental results demonstrate the efficiency of the proposed algorithm and the excellent predictive accuracy achieved by the selected model both in point and in probabilistic forecasting of the energy consumption.

Federated Learning (FL) is a paradigm for jointly training machine learning algorithms in a decentralized manner which allows for parties to communicate with an aggregator to create and train a model, without exposing the underlying raw data distribution of the local parties involved in the training process. Most research in FL has been focused on Neural Network-based approaches, however Tree-Based methods, such as XGBoost, have been underexplored in Federated Learning due to the challenges in overcoming the iterative and additive characteristics of the algorithm. Decision tree-based models, in particular XGBoost, can handle non-IID data, which is significant for algorithms used in Federated Learning frameworks since the underlying characteristics of the data are decentralized and have risks of being non-IID by nature. In this paper, we focus on investigating the effects of how Federated XGBoost is impacted by non-IID distributions by performing experiments on various sample size-based data skew scenarios and how these models perform under various non-IID scenarios. We conduct a set of extensive experiments across multiple different datasets and different data skew partitions. Our experimental results demonstrate that despite the various partition ratios, the performance of the models stayed consistent and performed close to or equally well against models that were trained in a centralized manner.

Many cyberattacks start with disseminating phishing URLs. When clicking these phishing URLs, the victim's private information is leaked to the attacker. There have been proposed several machine learning methods to detect phishing URLs. However, it still remains under-explored to detect phishing URLs with evasion, i.e., phishing URLs that pretend to be benign by manipulating patterns. In many cases, the attacker i) reuses prepared phishing web pages because making a completely brand-new set costs non-trivial expenses, ii) prefers hosting companies that do not require private information and are cheaper than others, iii) prefers shared hosting for cost efficiency, and iv) sometimes uses benign domains, IP addresses, and URL string patterns to evade existing detection methods. Inspired by those behavioral characteristics, we present a network-based inference method to accurately detect phishing URLs camouflaged with legitimate patterns, i.e., robust to evasion. In the network approach, a phishing URL will be still identified as phishy even after evasion unless a majority of its neighbors in the network are evaded at the same time. Our method consistently shows better detection performance throughout various experimental tests than state-of-the-art methods, e.g., F-1 of 0.89 for our method vs. 0.84 for the best feature-based method.

With the fast adoption of machine learning (ML) techniques, sharing of ML models is becoming popular. However, ML models are vulnerable to privacy attacks that leak information about the training data. In this work, we focus on a particular type of privacy attacks named property inference attack (PIA) which infers the sensitive properties of the training data through the access to the target ML model. In particular, we consider Graph Neural Networks (GNNs) as the target model, and distribution of particular groups of nodes and links in the training graph as the target property. While the existing work has investigated PIAs that target at graph-level properties, no prior works have studied the inference of node and link properties at group level yet. In this work, we perform the first systematic study of group property inference attacks (GPIA) against GNNs. First, we consider a taxonomy of threat models under both black-box and white-box settings with various types of adversary knowledge, and design six different attacks for these settings. We evaluate the effectiveness of these attacks through extensive experiments on three representative GNN models and three real-world graphs. Our results demonstrate the effectiveness of these attacks whose accuracy outperforms the baseline approaches. Second, we analyze the underlying factors that contribute to GPIA's success, and show that the target model trained on the graphs with or without the target property represents some dissimilarity in model parameters and/or model outputs, which enables the adversary to infer the existence of the property. Further, we design a set of defense mechanisms against the GPIA attacks, and demonstrate that these mechanisms can reduce attack accuracy effectively with small loss on GNN model accuracy.

In this article, we describe a {\tt R} package for sampling from an empirical likelihood-based posterior using a Hamiltonian Monte Carlo method. Empirical likelihood-based methodologies have been used in Bayesian modeling of many problems of interest in recent times. This semiparametric procedure can easily combine the flexibility of a non-parametric distribution estimator together with the interpretability of a parametric model. The model is specified by estimating equations-based constraints. Drawing an inference from a Bayesian empirical likelihood (BayesEL) posterior is challenging. The likelihood is computed numerically, so no closed expression of the posterior exists. Moreover, for any sample of finite size, the support of the likelihood is non-convex, which hinders the fast mixing of many Markov Chain Monte Carlo (MCMC) procedures. It has been recently shown that using the properties of the gradient of log empirical likelihood, one can devise an efficient Hamiltonian Monte Carlo (HMC) algorithm to sample from a BayesEL posterior. The package requires the user to specify only the estimating equations, the prior, and their respective gradients. An MCMC sample drawn from the BayesEL posterior of the parameters, with various details required by the user is obtained.

We consider the identification of large-scale linear and stable dynamic systems whose outputs may be the result of many correlated inputs. Hence, severe ill-conditioning may affect the estimation problem. This is a scenario often arising when modeling complex physical systems given by the interconnection of many sub-units where feedback and algebraic loops can be encountered. We develop a strategy based on Bayesian regularization where any impulse response is modeled as the realization of a zero-mean Gaussian process. The stable spline covariance is used to include information on smooth exponential decay of the impulse responses. We then design a new Markov chain Monte Carlo scheme that deals with collinearity and is able to efficiently reconstruct the posterior of the impulse responses. It is based on a variation of Gibbs sampling which updates possibly overlapping blocks of the parameter space on the basis of the level of collinearity affecting the different inputs. Numerical experiments are included to test the goodness of the approach where hundreds of impulse responses form the system and inputs correlation may be very high.

Markov chain Monte Carlo (MCMC) methods are frequently employed to sample from the posterior distribution of the parameters of interest. Such difficulties have restricted the use of Bayesian empirical likelihood (BayesEL) based methods in many applications. In this article, we propose a two-step Metropolis Hastings algorithm to sample from the BayesEL posteriors. Our proposal is specified hierarchically, where the estimating equations determining the empirical likelihood are used to propose values of a set of parameters depending on the proposed values of the remaining parameters. Furthermore, we discuss Bayesian model selection using empirical likelihood and extend our two-step Metropolis Hastings algorithm to a reversible jump Markov chain Monte Carlo procedure to sample from the resulting posterior. Finally, several applications of our proposed methods are presented. In recent years, empirical likelihood (Owen, 1988; Qin & Lawless, 1994) based procedures have been frequently used under Bayesian framework. Such procedures specify a statistical model through unbiased estimating equations, without requiring a declaration of the data distribution. The likelihood is estimated from the empirical distribution function computed under constraints imposed by these estimating equations. The estimated likelihood is then used to define a posterior. The validity of empirical and similar likelihoods for Bayesian inference has been a topic of extensive discussion (Monahan & Boos, 1992; Lazar, 2003; Fang & Mukerjee, 2006; Corcoran, 1998). Alternative likelihoods like Bayesian exponential tilted empirical likelihood (BETEL) (Schennach, 2005) have been proposed and justified using basic probabilistic arguments.