Learn More
This paper presents Recurrent Policy Gradients, a model-free reinforcement learning (RL) method creating limited-memory sto-chastic policies for partially observable Markov decision problems (POMDPs) that require long-term memories of past observations. The approach involves approximating a policy gradient for a Recurrent Neural Network (RNN) by(More)
Reinforcement learning for partially observable Markov decision problems (POMDPs) is a challenge as it requires policies with an internal state. Traditional approaches suffer significantly from this shortcoming and usually make strong assumptions on the problem domain such as perfect system models, state-estimators and a Markovian hidden system. Recurrent(More)
We present curiosity-driven, autonomous acquisition of tactile exploratory skills on a biomimetic robot finger equipped with an array of microelectromechanical touch sensors. Instead of building tailored algorithms for solving a specific tactile task, we employ a more general curiosity-driven reinforcement learning approach that autonomously learns a set of(More)
We describe a new algorithm for robot localization, efficient both in terms of memory and processing time. It transforms a stream of laser range sensor data into a probabilistic calculation of the robot's position , using a bidirectional Long Short-Term Memory (LSTM) recurrent neural network (RNN) to learn the structure of the environment and to answer(More)
—In this paper, we propose a novel architecture for wireless sensor network testbeds, called MOTEL. The main novelty compared to existing architectures is the possibility to include mobile sensor nodes. To support mobility, we deal with two main challenges: controlled mobility of sensor nodes, and the need to operate sensor nodes in the absence of a(More)
Business-driven development favors the construction of process models at different abstraction levels and by different people. As a consequence, there is a demand for consolidating different versions of process models by detecting and resolving differences. Existing approaches rely on the existence of a change log which logs the changes when changing a(More)
Business processes usually have to consider certain constraints like domain specific and quality requirements. The automated formal verification of these constraints is desirable , but requires the user to provide an unambiguous formal specification. In particular since the notations for business process modeling are usually visual flow-oriented languages,(More)