Publications (18)
The Effectiveness of Discretization in Forecasting: An Empirical Study on Neural Time Series Models
Stephan Rabanser, Tim Januschowski, Valentin Flunkert +2
Time series modeling techniques based on deep learning have seen many advancements in recent years, especially in data-abundant settings and with the central aim of learning global…
Intrinsic Anomaly Detection for Multi-Variate Time Series
Stephan Rabanser, Tim Januschowski, Kashif Rasul +6
We introduce a novel, practically relevant variation of the anomaly detection problem in multi-variate time series: intrinsic anomaly detection. It appears in diverse practical sce…
Approximate Bayesian Inference in Linear State Space Models for Intermittent Demand Forecasting at Scale
Matthias Seeger, Syama Rangapuram, Yuyang Wang +4
We present a scalable and robust Bayesian inference method for linear state space models. The method is applied to demand forecasting in the context of a large e-commerce platform,…
DeepAR: Probabilistic Forecasting with Autoregressive Recurrent Networks
David Salinas, Valentin Flunkert, Jan Gasthaus
Probabilistic forecasting, i.e. estimating the probability distribution of a time series' future given its past, is a key enabler for optimizing business processes. In retail busin…
Symmetry-breaking transitions in networks of nonlinear circuit elements
Martin Heinrich, Thomas Dahms, Valentin Flunkert +2
We investigate a nonlinear circuit consisting of N tunnel diodes in series, which shows close similarities to a semiconductor superlattice or to a neural network. Each tunnel diode…
Discontinuous Attractor Dimension at the Synchronization Transition of Time-Delayed Chaotic Systems
Steffen Zeeb, Thomas Dahms, Valentin Flunkert +3
The attractor dimension at the transition to complete synchronization in a network of chaotic units with time-delayed couplings is investigated. In particular, we determine the Kap…
Deep Learning for Time Series Forecasting: Tutorial and Literature Survey
Konstantinos Benidis, Syama Sundar Rangapuram, Valentin Flunkert +10
Deep learning based forecasting methods have become the methods of choice in many applications of time series prediction or forecasting often outperforming other approaches. Conseq…
A simple and effective predictive resource scaling heuristic for large-scale cloud applications
Valentin Flunkert, Quentin Rebjock, Joel Castellon +2
We propose a simple yet effective policy for the predictive auto-scaling of horizontally scalable applications running in cloud environments, where compute resources can only be ad…
Deep Non-Parametric Time Series Forecaster
Syama Sundar Rangapuram, Jan Gasthaus, Lorenzo Stella +4
This paper presents non-parametric baseline models for time series forecasting. Unlike classical forecasting models, the proposed approach does not assume any parametric form for t…
Neural Contextual Anomaly Detection for Time Series
Chris U. Carmona, François-Xavier Aubet, Valentin Flunkert +1
We introduce Neural Contextual Anomaly Detection (NCAD), a framework for anomaly detection on time series that scales seamlessly from the unsupervised to supervised setting, and is…
Adaptive Tuning of Feedback Gain in Time-Delayed Feedback Control
Judith Lehnert, Philipp Hövel, Valentin Flunkert +3
We demonstrate that time-delayed feedback control can be improved by adaptively tuning the feedback gain. This adaptive controller is applied to the stabilization of an unstable fi…
Meta-Forecasting by combining Global Deep Representations with Local Adaptation
Riccardo Grazzi, Valentin Flunkert, David Salinas +3
While classical time series forecasting considers individual time series in isolation, recent advances based on deep learning showed that jointly learning from a large pool of rela…
Strong and weak chaos in nonlinear networks with time-delayed couplings
Sven Heiligenthal, Thomas Dahms, Serhiy Yanchuk +5
We study chaotic synchronization in networks with time-delayed coupling. We introduce the notion of strong and weak chaos, distinguished by the scaling properties of the maximum Ly…
GluonTS: Probabilistic Time Series Models in Python
Alexander Alexandrov, Konstantinos Benidis, Michael Bohlke-Schneider +10
We introduce Gluon Time Series (GluonTS, available at https://gluon-ts.mxnet.io), a library for deep-learning-based time series modeling. GluonTS simplifies the development of and…
Chaos synchronization in networks of delay-coupled lasers: Role of the coupling phases
Valentin Flunkert, Eckehard Schöll
We derive rigorous conditions for the synchronization of all-optically coupled lasers. In particular, we elucidate the role of the optical coupling phases for synchronizability by…
Criteria for Classifying Forecasting Methods
Tim Januschowski, Jan Gasthaus, Yuyang Wang +4
Classifying forecasting methods as being either of a "machine learning" or "statistical" nature has become commonplace in parts of the forecasting literature and community, as exem…
Suppressing noise-induced intensity pulsations in semiconductor lasers by means of time-delayed feedback
Valentin Flunkert, Eckehard Schoell
We investigate the possibility to suppress noise-induced intensity pulsations (relaxation oscillations) in semiconductor lasers by means of a time-delayed feedback control scheme.…
Multi-Objective Model Selection for Time Series Forecasting
Oliver Borchert, David Salinas, Valentin Flunkert +2
Research on time series forecasting has predominantly focused on developing methods that improve accuracy. However, other criteria such as training time or latency are critical in…