Machine learning algorithms have been shown to be vulnerable to adversarial manipulation through systematic modification of inputs (e.g., adversarial examples) in domains such as image recognition. Under the default threat model, the adversary exploits the unconstrained nature of images; each feature (pixel) is fully under control of the adversary. However, it is not clear how these attacks translate to constrained domains that limit which and how features can be modified by the adversary (e.g., network intrusion detection). In this paper, we explore whether constrained domains are less vulnerable than unconstrained domains to adversarial example generation algorithms. We create an algorithm for generating adversarial sketches: targeted universal perturbation vectors which encode feature saliency within the envelope of domain constraints. To assess how these algorithms perform, we evaluate them in constrained (e.g., network intrusion detection) and unconstrained (e.g., image recognition) domains. The results demonstrate that our approaches generate misclassification rates in constrained domains that were comparable to those of unconstrained domains (greater than 95%). Our investigation shows that the narrow attack surface exposed by constrained domains is still sufficiently large to craft successful adversarial examples; and thus, constraints do not appear to make a domain robust. Indeed, with as little as five randomly selected features, one can still generate adversarial examples.

In this paper, we propose a new paradigm for paraphrase generation by treating the task as unsupervised machine translation (UMT) based on the assumption that there must be pairs of sentences expressing the same meaning in a large-scale unlabeled monolingual corpus. The proposed paradigm first splits a large unlabeled corpus into multiple clusters, and trains multiple UMT models using pairs of these clusters. Then based on the paraphrase pairs produced by these UMT models, a unified surrogate model can be trained to serve as the final \sts model to generate paraphrases, which can be directly used for test in the unsupervised setup, or be finetuned on labeled datasets in the supervised setup. The proposed method offers merits over machine-translation-based paraphrase generation methods, as it avoids reliance on bilingual sentence pairs. It also allows human intervene with the model so that more diverse paraphrases can be generated using different filtering criteria. Extensive experiments on existing paraphrase dataset for both the supervised and unsupervised setups demonstrate the effectiveness the proposed paradigm.

Working memory (WM), denoting the information temporally stored in the mind, is a fundamental research topic in the field of human cognition. Electroencephalograph (EEG), which can monitor the electrical activity of the brain, has been widely used in measuring the level of WM. However, one of the critical challenges is that individual differences may cause ineffective results, especially when the established model meets an unfamiliar subject. In this work, we propose a cross-subject deep adaptation model with spatial attention (CS-DASA) to generalize the workload classifications across subjects. First, we transform EEG time series into multi-frame EEG images incorporating spatial, spectral, and temporal information. First, the Subject-Shared module in CS-DASA receives multi-frame EEG image data from both source and target subjects and learns the common feature representations. Then, in the subject-specific module, the maximum mean discrepancy is implemented to measure the domain distribution divergence in a reproducing kernel Hilbert space, which can add an effective penalty loss for domain adaptation. Additionally, the subject-to-subject spatial attention mechanism is employed to focus on the discriminative spatial features from the target image data. Experiments conducted on a public WM EEG dataset containing 13 subjects show that the proposed model is capable of achieving better performance than existing state-of-the-art methods.

When you're working on a machine learning project, the most tedious steps are often data cleaning and preprocessing. Especially when you're working in a Jupyter Notebook, running code in many cells can be confusing. The Scikit-learn library has tools called Pipeline and ColumnTransformer that can really make your life easier. Instead of transforming the dataframe step by step, the pipeline combines all transformation steps. You can get the same result with less code.

There are many R packages for working with Time Series data. Here's how timetk compares to the "tidy" time series R packages for data visualization, wrangling, and feature engineeering (those that leverage data frames or tibbles). Timetk is an amazing package that is part of the modeltime ecosystem for time series analysis and forecasting. Your probably thinking how am I ever going to learn time series forecasting. Here's the solution that will save you years of struggling.

In simpler terms, it is the quantization of color spaces. Color spaces are a way to characterize the shade channels existing in the photo that offers the photograph that precise hue. This is a useful image compression technique which is quite useful for devices that can show a limited number of colors due to memory restriction. Each image can be represented by three features: the R, G and B values for each pixel. Given that our image has pixel values ranging from 0 to 255, we can say that each image has 256 * 256 * 256 colors. Our goal now is to reduce the number of colors to a manageable number.

Asthma is a common chronic disease of the respiratory system causing significant disability and societal burden. It affects more than 300 million people worldwide, while more than 100 million people will likely have asthma by 2025. The price of asthma varies greatly from nation to nation. Mean yearly cost can be estimated to 1900 EUR in Europe and $3100 in the United States. Managing asthma involves controlling symptoms, preventing exacerbations, and maintaining lung function. Improved asthma control is reduces the risk of exacerbations and lung function impairment while reducing the direct costs of asthma care and indirect costs associated with reduced productivity. Understanding the complex dynamics of the pulmonary system and the lung's response to disease is fundamental to the advancement of Asthma treatment. Computational models of the respiratory system seek to provide a theoretical framework to understand the interaction between structure and function. Their application can improve pulmonary medicine by a patient-specific approach to medicinal methodologies optimizing the delivery given the personalized geometry and personalized ventilation patterns. A three-fold objective is addressed within this dissertation. The first part refers to the comprehension of pulmonary pathophysiology and the mechanics of Asthma and subsequently of constrictive pulmonary conditions in general. The second part refers to the design and implementation of tools that facilitate personalized medicine to improve delivery and effectiveness. Finally, the third part refers to the self-management of the condition, meaning that medical personnel and patients have access to tools and methods that allow the first party to easily track the course of the condition and the second party, i.e. the patient to easily self-manage it alleviating the significant burden from the health system.

A ubiquitous problem in machine learning, statistics, and signal processing is to accurately estimate an unknown sparse vector from a few noisy linear measurements. This estimation problem, which we refer to as sparse coding, is at the heart of the field of compressed sensing, revealing that under sparsity assumptions it is possible to successfully recover a signal that sampled significantly below the Nyquist rate [1, 2]. This, in turn, led to a dramatic increase in magnetic resonance imaging (MRI) scanning session speed [3]. Another exciting application that also builds on the sparsity assumption is unsupervised representation learning, i.e., given high-dimensional input data, such as an image, finding a low-dimensional representation that captures the intrinsic underlying structure in the input [4, 5, 6]. These representations are often used in image restoration tasks to effectively remove noise (denoising) [7, 8], fill-in missing pixels (inpainting) [9, 10, 11], and to achieve high quality digital zoom (super-resolution) [10, 12, 13, 14]. Sparsity also plays a key role in linear regression when given a large pool of features, to form a predictive rule that estimates an unknown response using a smaller, interpretable subset of features that manifests the strongest effects [15, 16, 17, 18]. To formalize the sparse coding problem, which is central for tackling the aforementioned applications, we consider the following linear model: b = Ax + v, where A is a matrix of size M N, the vector x is of length N, and v is a noise vector of length M. In this paper, we focus on a challenging setting in which M N, where a crucial assumption we make is that the vector x is k-sparse, i.e., it contains only k non-zero elements with k N [2, 1, 19].

Privacy-preserving data analysis has become prevailing in recent years. In this paper, we propose a distributed group differentially private majority vote mechanism for the sign selection problem in a distributed setup. To achieve this, we apply the iterative peeling to the stability function and use the exponential mechanism to recover the signs. As applications, we study the private sign selection for mean estimation and linear regression problems in distributed systems. Our method recovers the support and signs with the optimal signal-to-noise ratio as in the non-private scenario, which is better than contemporary works of private variable selections. Moreover, the sign selection consistency is justified with theoretical guarantees. Simulation studies are conducted to demonstrate the effectiveness of our proposed method.

Subsampling or subdata selection is a useful approach in large-scale statistical learning. Most existing studies focus on model-based subsampling methods which significantly depend on the model assumption. In this paper, we consider the model-free subsampling strategy for generating subdata from the original full data. In order to measure the goodness of representation of a subdata with respect to the original data, we propose a criterion, generalized empirical F-discrepancy (GEFD), and study its theoretical properties in connection with the classical generalized L2-discrepancy in the theory of uniform designs. These properties allow us to develop a kind of low-GEFD data-driven subsampling method based on the existing uniform designs. By simulation examples and a real case study, we show that the proposed subsampling method is superior to the random sampling method. Moreover, our method keeps robust under diverse model specifications while other popular subsampling methods are under-performing. In practice, such a model-free property is more appealing than the model-based subsampling methods, where the latter may have poor performance when the model is misspecified, as demonstrated in our simulation studies.