Harald Burgsteiner

Movement prediction from real-world images using a liquid state machine

Prediction is an important task in robot motor control where it is used to gain feedback for a controller. With such a self-generated feedback, which is available before sensor readings from an environment can be processed, a controller can be stabilized and thus the performance of a moving robot in a real-world environment is improved. So far, only experiments with artificially generated data have shown good results. In a sequence of experiments we evaluate whether a liquid state machine in combination with a supervised learning algorithm can be used to predict ball trajectories with input data coming from a video camera mounted on a robot participating in the RoboCup. This pre-processed video data is fed into a recurrent spiking neural network. Connections to some output neurons are trained by linear regression to predict the position of a ball in various time steps ahead. Our results support the idea that learning with a liquid state machine can be applied not only to designed data but also to real, noisy data.

Reducing Communication for Distributed Learning in Neural Networks

A learning algorithm is presented for circuits consisting of a single layer of perceptrons. We refer to such circuits as parallel perceptrons. In spite of their simplicity,these circuits are universal approximators for arbitrary boolean and continuous functions. In contrast to backprop for multi-layer perceptrons,our new learning algorithm - the parallel delta rule (p-delta rule) - only has to tune a single layer of weights,and it does not require the computation and communication of analog values with high precision. Reduced communication also distinguishes our new learning rule from other learning rules for such circuits such as those traditionally used for MADALINE. A theoretical analysis shows that the p-delta rule does in fact implement gradient descent - with regard to a suitable error measure - although it does not require to compute derivatives. Furthermore it is shown through experiments on common real-world benchmark datasets that its performance is competitive with that of other learning approaches from neural networks and machine learning. Thus our algorithm also provides an interesting new hypothesis for the organization of learning in biological neural systems.

Imitation learning with spiking neural networks and real-world devices

This article is about a new approach in robotic learning systems. It provides a method to use a real-world device that operates in real-time, controlled through a simulated recurrent spiking neural network for robotic experiments. A randomly generated network is used as the main computational unit. Only the weights of the output units of this network are changed during training. It will be shown, that this simple type of a biological realistic spiking neural network-also known as a neural microcircuit-is able to imitate robot controllers like that incorporated in Braitenberg vehicles. A more non-linear type controller is imitated in a further experiment. In a different series of experiments that involve temporal memory reported in Burgsteiner et al. [2005. In: Proceedings of the 18th International Conference IEA/AIE. Lecture Notes in Artificial Intelligence. Springer, Berlin, pp. 121-130.] this approach also provided a basis for a movement prediction task. The results suggest that a neural microcircuit with a simple learning rule can be used as a sustainable robot controller for experiments in computational motor control.

Robust Bone Marrow Cell Discrimination by Rotation-Invariant Training of Multi-class Echo State Networks

Classification of cell types in context of the architecture in tissue specimen is the basis of diagnostic pathology and decisions for comprehensive investigations rely on a valid interpretation of tissue morphology. Especially visual examination of bone marrow cells takes a considerable amount of time and inter-observer variability can be remarkable. In this work, we propose a novel rotation-invariant learning scheme for multi-class Echo State Networks (ESNs), which achieves very high performance in automated bone marrow cell classification. Based on representing static images as temporal sequence of rotations, we show how ESNs robustly recognize cells of arbitrary rotations by taking advantage of their short-term memory capacity.

Training echo state networks for rotation-invariant bone marrow cell classification

The main principle of diagnostic pathology is the reliable interpretation of individual cells in context of the tissue architecture. Especially a confident examination of bone marrow specimen is dependent on a valid classification of myeloid cells. In this work, we propose a novel rotation-invariant learning scheme for multi-class echo state networks (ESNs), which achieves very high performance in automated bone marrow cell classification. Based on representing static images as temporal sequence of rotations, we show how ESNs robustly recognize cells of arbitrary rotations by taking advantage of their short-term memory capacity. The performance of our approach is compared to a classification random forest that learns rotation-invariance in a conventional way by exhaustively training on multiple rotations of individual samples. The methods were evaluated on a human bone marrow image database consisting of granulopoietic and erythropoietic cells in different maturation stages. Our ESN approach to cell classification does not rely on segmentation of cells or manual feature extraction and can therefore directly be applied to image data.

Improving emergency medical dispatching by means of the Adaptive Large Neighborhood Search

Dispatching ambulances is a demanding and stressful task for dispatchers. The aim of this work was to investigate if and to what extent the dispatch operation of the Red Cross Salzburg (RCS) can be optimized with a computerized system. The basic problem of a dynamic multi-vehicle Dial-a-Ride Problem (DARP) with time windows was enhanced according to the requirements of the RCS. The objective was to minimize the total mileage covered by ambulances and the waiting time of patients. Consequently, the problem was solved by using the Adaptive Large Neighborhood Search (ALNS). Evaluation results indicate that the system outperforms a human dispatcher by between 2.5% and 36% within 1 minute of runtime.