Frequency-based Feature Selection for Imbalanced Time-Series Data – Learning a representation of a data set can greatly simplify the annotation of the data using sparsely sampled samples. In this paper, we present a novel clustering-based approach based on the principle of minimizing the maximum likelihood minimization (MLE). Here, the MLE is defined as a linear family of estimators that is equivalent to the maximum likelihood minimization (LFN) of a set. We show how to build a model that maps the MLE to a subset of the data, and compare to LFN for the case of a sparsely sampled set. Experimental results show that the proposed framework outperforms the LFN estimators, providing a new approach for inference based on information extraction. The model can be constructed as a graph from a sparse set of data.
In some applications, a data-dependent representation of the data may provide insights into the distribution of uncertainty associated with the measurement error. Such an insight can be used in a variety of applications, such as learning to predict a given event, learning from noisy measurements of the underlying structure in a given data, and learning to predict a given distribution of uncertainty at a target time. In this paper, we propose a novel Bayesian inference framework to obtain predictive distributions over the observed data. The proposed framework relies on a Bayesian approach to inference in unstructured data, where only the observed data are available. We provide an efficient method for inference, provide a Bayesian framework to optimize a model, and demonstrate both the utility of the Bayesian framework and the ability to leverage the uncertainty as inputs to the inference task.
Automated Evaluation of Neural Networks for Polish Machine-Patch Recognition
Deep Convolutional Neural Network for Brain Decoding
Frequency-based Feature Selection for Imbalanced Time-Series Data
Single image super resolution with the maximum density embedding prichon linear model
Learning Disentangled Representations with Latent Factor ModelingIn some applications, a data-dependent representation of the data may provide insights into the distribution of uncertainty associated with the measurement error. Such an insight can be used in a variety of applications, such as learning to predict a given event, learning from noisy measurements of the underlying structure in a given data, and learning to predict a given distribution of uncertainty at a target time. In this paper, we propose a novel Bayesian inference framework to obtain predictive distributions over the observed data. The proposed framework relies on a Bayesian approach to inference in unstructured data, where only the observed data are available. We provide an efficient method for inference, provide a Bayesian framework to optimize a model, and demonstrate both the utility of the Bayesian framework and the ability to leverage the uncertainty as inputs to the inference task.