Markus Schneider

University of Applied Sciences Ravensburg-Weingarten | RWU · Machine Learning

Prostate segmentation is a substantial factor in the diagnostic pathway of suspicious prostate lesions. Medical doctors are assisted by computer-aided detection and diagnosis systems systems and methods derived from artificial intelligence deep learning-based systems have to be trained on existing data. Especially freely available labeled prostate...

Reliable tree log detection is a key requirement for automation of forestry operations. Despite the substantial progress regarding object detection in general, tree log detection lags behind due to the lack of well-annotated datasets. In order to address this gap, we introduce the Tree Log Detection Dataset (TLDD). This real-world dataset is collec...

We consider the problem of univariate time series prediction from an elementary machine learning point of view. Beginning with the question of whether and how Principal Component Analysis (PCA) can be used for time series prediction, we describe a simple methodology and attempt to classify PCA-based prediction in terms of statistics, signal process...

Many traditional machine learning and pattern recognition algorithms—as for example linear discriminant analysis (LDA) or principal component analysis (PCA)—optimize data representation with respect to an information theoretic criterion. For time series analysis these traditional techniques are typically insufficient. In this work we propose an ext...

We present a novel approach to spectral estimation, posing power spectrum density (PSD) estimation as a constrained optimization problem. More precisely, weoptimize the coefficients of a finite impulse response filter such that the output variance is maximized. An analytical solution for this optimization problem is provided and we show that its so...

Singular Spectrum Analysis (SSA) is a powerful method that is frequently used in dynamical systems theory and time series analysis. However, the algorithm itself is only partially understood. In this paper, we tackle the problem of a thorough interpretation of the complete basic SSA algorithm. We point out the relationship between SSA and Fourier a...

Principal component analysis (PCA) and kernel PCA allow the decorrelation of data with respect to a basis that is found via variance maximization. However, these techniques are based on pointwise correlations. Especially in the context of time series analysis this is not optimal. We present a novel generalization of PCA that allows to imprint any d...

Principal component analysis (PCA), a well-known technique in machine learning and statistics, is typically applied to time-independent data, as it is based on point-wise correlations. Dynamic PCA (DPCA) handles this issue by augmenting the data set with lagged versions of itself. In this paper, we show that both, PCA and DPCA, are a special case o...

We present a novel algorithm for anomaly detection on very large datasets and data streams. The method, named EXPected Similarity Estimation (EXPoSE), is kernel-based and able to efficiently compute the similarity between new data points and the distribution of regular data. The estimator is formulated as an inner product with a reproducing kernel...

We present an approach for large-scale batch and streaming anomaly detection based on the kernel embedding of distributions named EXPected Similarity Estimation (EXPoSE). This algorithm has a constant memory consumption, requires constant time per prediction and can be trained incrementally with a computational complexity which scales linear with t...

The kernel embedding of distributions is a popular machine learning technique to manipulate probability distributions and is an integral part of numerous applications. Its empirical counterpart is an estimate from a finite set of samples from the distribution under consideration. However, for large-scale learning problems the empirical kernel embed...

A new algorithm named EXPected Similarity Estimation (EXPoSE) was recently proposed to solve the problem of large-scale anomaly detection. It is a non-parametric and distribution free kernel method based on the Hilbert space embedding of probability measures. Given a dataset of $n$ samples, EXPoSE needs only $\mathcal{O}(n)$ (linear time) to build...

The approximation of kernel functions using explicit feature maps gained a lot of attention in recent years due to the tremendous speed up in training and learning time of kernel-based algorithms, making them applicable to very large-scale problems. For example, approximations based on random Fourier features are an efficient way to create feature...

We propose a new algorithm named EXPected Similarity Estimation (EXPoSE) to approach the problem of anomaly detection (also known as one-class learning or outlier detection) which is based on the similarity between data points and the distribution of non-anomalous data. We formulate the problem as an inner product in a reproducing kernel Hilbert sp...

Learning from Demonstration (LfD) is a powerful method for training robots to solve tasks involving low level motion skills, thus avoiding human programming effort. We present Learning from Demonstration by Averaging Trajecto- ries (LAT) which is a new, simple and computationally fast method and provide an implementation on a service robot. We comp...

We tackle the problem of multi-task learning with copula process. Multivariable prediction in spatial and spatial-temporal processes such as natural resource estimation and pollution monitoring have been typically addressed using techniques based on Gaussian processes and co-Kriging. While the Gaussian prior assumption is convenient from analytical...

In recent years there was a tremendous progress in robotic systems, and however also increased expectations: A robot should be easy to program and reliable in task execution. Learning from Demonstration (LfD) offers a very promising alternative to classical engineering approaches. LfD is a very natural way for humans to interact with robots and wil...

Today, robots are already able to solve specific tasks in laboratory environments. Since everyday environments are more complex, the robot skills required to solve everyday tasks cannot be known in advance and thus not be programmed beforehand. Rather, the robot must be able to learn those tasks being instructed by users without any technical backg...