We propose cKAM, cyclical Kernel Adaptive Metropolis, which incorporates a cyclical stepsize scheme to allow control for exploration and sampling. We show that on a crafted bimodal distribution, existing Adaptive Metropolis type algorithms would fail to converge to the true posterior distribution. We point out that this is because adaptive samplers estimates the local/global covariance structure using past history of the chain, which will lead to adaptive algorithms be trapped in a local mode. We demonstrate that cKAM encourages exploration of the posterior distribution and allows the sampler to escape from a local mode, while maintaining the high performance of adaptive methods.

Users have the right to have their data deleted by third-party learned systems, as codified by recent legislation such as the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA). Such data deletion can be accomplished by full re-training, but this incurs a high computational cost for modern machine learning models. To avoid this cost, many approximate data deletion methods have been developed for supervised learning. Unsupervised learning, in contrast, remains largely an open problem when it comes to (approximate or exact) efficient data deletion. In this paper, we propose a density-ratio-based framework for generative models. Using this framework, we introduce a fast method for approximate data deletion and a statistical test for estimating whether or not training points have been deleted. We provide theoretical guarantees under various learner assumptions and empirically demonstrate our methods across a variety of generative methods.

More robust approaches are studied to this effect Data smoothing is heavily utilized in many learning [41]-[43], where a prominent example is the moving median problems to detect certain patters in the data or mitigate filter. Although the moving median is a nonlinear smoother noisy/anomalous values [1]. The notion of smoothing stems and the smoothed signal is acquired by the weighted median from the inference that the data points close to each other in of close groups of samples, it is efficiently computable by their features should have close mappings (functional evaluations), use of indexable skiplist [44]. We point out that specifically i.e., behave similarly. The most prominent example is for the moving median, the smoothed signal is statistically a time-series dataset where data points closer in time should optimal in the maximum likelihood sense, when the signal have similar values, e.g, unnaturally high or low values should contamination results from a Laplace distribution [45]. The be suppressed to match the nominal data pattern. Smoothing moving median filter has many applications including mass is mainly done on the datasets for easier extraction of their spectrometry [46], edge detection [47]. The application is information and consequently ease their analyses [2]. This especially rich in the field of image processing because of approach is applied in many fields including signal processing the edge-preserving capabilities [48].

In learning from label proportions (LLP), the instances are grouped into bags, and the task is to learn an instance classifier given relative class proportions in training bags. LLP is useful when obtaining individual instance labels is impossible or costly. In this work, we focus on the case of small bags, which allows to design an algorithm that explicitly considers all consistent instance label combinations. In particular, we propose an EM algorithm alternating between optimizing a general neural network instance classifier and incorporating bag-level annotations. Using two different image datasets, we experimentally compare this method with an approach based on normal approximation and two existing LLP methods. The results show that our approach converges faster to a comparable or better solution.

Human speakers encode information into raw speech which is then decoded by the listeners. This complex relationship between encoding (production) and decoding (perception) is often modeled separately. Here, we test how encoding and decoding of lexical semantic information can emerge automatically from raw speech in unsupervised generative deep convolutional networks that combine the production and perception principles of speech. We introduce, to our knowledge, the most challenging objective in unsupervised lexical learning: a network that must learn unique representations for lexical items with no direct access to training data. We train several models (ciwGAN and fiwGAN arXiv:2006.02951) and test how the networks classify acoustic lexical items in unobserved test data. Strong evidence in favor of lexical learning and a causal relationship between latent codes and meaningful sublexical units emerge. The architecture that combines the production and perception principles is thus able to learn to decode unique information from raw acoustic data without accessing real training data directly. We propose a technique to explore lexical (holistic) and sublexical (featural) learned representations in the classifier network. The results bear implications for unsupervised speech technology, as well as for unsupervised semantic modeling as language models increasingly bypass text and operate from raw acoustics.

Machine Learning models are frequently seen as black boxes that are impossible to decipher. Because the learner is trained to respond to "yes" and "no" type questions without explaining how the answer was obtained. An explanation of how an answer was achieved is critical in many applications for assuring confidence and openness. Explainable AI refers to strategies and procedures in the use of artificial intelligence technology (AI) that allow human specialists to understand the solution's findings. This article will focus on explaining the machine learner using OmniXAI.

You are right that randomness will play a role (like with many other algorithms including MCMC samplers for Bayesian models, XGBoost, LightGBM, neural networks etc.) in the results. The obvious way to minimize randomness in the results of any hyper-parameter optimization method for RF (whether it's random grid-search, grid search or some Bayesian hyperparameter optimization method) is to increase the number of trees (which reduces the randomness in the model behavior - albeit at the cost of an increased training time). Alternatively, you construct a surrogate model on top of the results that takes into account that the signal, of where the best model in the hyperparameter landscape is, is noisy through an appropriate amount of smoothing/regularization.

Machine learning is rapidly becoming as ubiquitous as data itself. Quite literally wherever there is an abundance of data, machine learning is somehow intertwined. After all, what utility would data have if we were not able to use it to predict something about the future? Luckily there is a plethora of toolkits and frameworks that have made it rather simple to deploy ML in Python. Specifically, Sklearn has done a terrifically effective job at making ML accessible to developers.

Abstract: We study the Stochastic Gradient Descent (SGD) algorithm in nonparametric statistics: kernel regression in particular. The directional bias property of SGD, which is known in the linear regression setting, is generalized to the kernel regression. More specifically, we prove that SGD with moderate and annealing step-size converges along the direction of the eigenvector that corresponds to the largest eigenvalue of the Gram matrix. These facts are referred to as the directional bias properties; they may interpret how an SGD-computed estimator has a potentially smaller generalization error than a GD-computed estimator. The application of our theory is demonstrated by simulation studies and a case study that is based on the FashionMNIST dataset.

Forward modeling approaches in cosmology have made it possible to reconstruct the initial conditions at the beginning of the Universe from the observed survey data. However the high dimensionality of the parameter space still poses a challenge to explore the full posterior, with traditional algorithms such as Hamiltonian Monte Carlo (HMC) being computationally inefficient due to generating correlated samples and the performance of variational inference being highly dependent on the choice of divergence (loss) function. Here we develop a hybrid scheme, called variational self-boosted sampling (VBS) to mitigate the drawbacks of both these algorithms by learning a variational approximation for the proposal distribution of Monte Carlo sampling and combine it with HMC. The variational distribution is parameterized as a normalizing flow and learnt with samples generated on the fly, while proposals drawn from it reduce auto-correlation length in MCMC chains. Our normalizing flow uses Fourier space convolutions and element-wise operations to scale to high dimensions. We show that after a short initial warm-up and training phase, VBS generates better quality of samples than simple VI approaches and reduces the correlation length in the sampling phase by a factor of 10-50 over using only HMC to explore the posterior of initial conditions in 64$^3$ and 128$^3$ dimensional problems, with larger gains for high signal-to-noise data observations.