Search Results for author: Thomas Runkler

Found 25 papers, 7 papers with code

User-Interactive Offline Reinforcement Learning

no code implementations21 May 2022 Phillip Swazinna, Steffen Udluft, Thomas Runkler

At the same time, offline RL algorithms are not able to tune their most important hyperparameter - the proximity of the learned policy to the original policy.

Offline RL reinforcement-learning

How to Manage Tiny Machine Learning at Scale: An Industrial Perspective

1 code implementation18 Feb 2022 Haoyu Ren, Darko Anicic, Thomas Runkler

Tiny machine learning (TinyML) has gained widespread popularity where machine learning (ML) is democratized on ubiquitous microcontrollers, processing sensor data everywhere in real-time.

Comparing Model-free and Model-based Algorithms for Offline Reinforcement Learning

1 code implementation14 Jan 2022 Phillip Swazinna, Steffen Udluft, Daniel Hein, Thomas Runkler

Offline reinforcement learning (RL) Algorithms are often designed with environments such as MuJoCo in mind, in which the planning horizon is extremely long and no noise exists.

Offline RL reinforcement-learning

Combining Sub-Symbolic and Symbolic Methods for Explainability

no code implementations3 Dec 2021 Anna Himmelhuber, Stephan Grimm, Sonja Zillner, Mitchell Joblin, Martin Ringsquandl, Thomas Runkler

Similarly to other connectionist models, Graph Neural Networks (GNNs) lack transparency in their decision-making.

Decision Making

Measuring Data Quality for Dataset Selection in Offline Reinforcement Learning

no code implementations26 Nov 2021 Phillip Swazinna, Steffen Udluft, Thomas Runkler

Recently developed offline reinforcement learning algorithms have made it possible to learn policies directly from pre-collected datasets, giving rise to a new dilemma for practitioners: Since the performance the algorithms are able to deliver depends greatly on the dataset that is presented to them, practitioners need to pick the right dataset among the available ones.

reinforcement-learning

Ontology-Based Skill Description Learning for Flexible Production Systems

no code implementations25 Nov 2021 Anna Himmelhuber, Stephan Grimm, Thomas Runkler, Sonja Zillner

The increasing importance of resource-efficient production entails that manufacturing companies have to create a more dynamic production environment, with flexible manufacturing machines and processes.

Inductive logic programming

Demystifying Graph Neural Network Explanations

no code implementations25 Nov 2021 Anna Himmelhuber, Mitchell Joblin, Martin Ringsquandl, Thomas Runkler

Graph neural networks (GNNs) are quickly becoming the standard approach for learning on graph structured data across several domains, but they lack transparency in their decision-making.

Decision Making Synthetic Data Generation

Towards Data-Free Domain Generalization

no code implementations9 Oct 2021 Ahmed Frikha, Haokun Chen, Denis Krompaß, Thomas Runkler, Volker Tresp

We define the novel problem of Data-Free Domain Generalization (DFDG), a practical setting where models trained on the source domains separately are available instead of the original datasets, and investigate how to effectively solve the domain generalization problem in that case.

Domain Generalization Knowledge Distillation

Learning through structure: towards deep neuromorphic knowledge graph embeddings

1 code implementation21 Sep 2021 Victor Caceres Chian, Marcel Hildebrandt, Thomas Runkler, Dominik Dold

Over the recent years, a multitude of different graph neural network architectures demonstrated ground-breaking performances in many learning tasks.

Graph Learning Knowledge Graph Embedding +3

Behavior Constraining in Weight Space for Offline Reinforcement Learning

1 code implementation12 Jul 2021 Phillip Swazinna, Steffen Udluft, Daniel Hein, Thomas Runkler

In offline reinforcement learning, a policy needs to be learned from a single pre-collected dataset.

reinforcement-learning

The Synergy of Complex Event Processing and Tiny Machine Learning in Industrial IoT

no code implementations4 May 2021 Haoyu Ren, Darko Anicic, Thomas Runkler

Focusing on comprehensive networking, big data, and artificial intelligence, the Industrial Internet-of-Things (IIoT) facilitates efficiency and robustness in factory operations.

TinyOL: TinyML with Online-Learning on Microcontrollers

no code implementations15 Mar 2021 Haoyu Ren, Darko Anicic, Thomas Runkler

The neural network is first trained using a large amount of pre-collected data on a powerful machine and then flashed to MCUs.

online learning

CLARE-GAN: GENERATION OF CLASS-SPECIFIC TIME SERIES

no code implementations1 Jan 2021 Hiba Arnout, Johanna Bronner, Thomas Runkler

We prove that our model outperforms the state-of-the-art generative models and leads to a significant and consistent improvement in the quality of the generated time series while at the same time preserving the classes and the variation of the original dataset.

Time Series

Overcoming Model Bias for Robust Offline Deep Reinforcement Learning

no code implementations12 Aug 2020 Phillip Swazinna, Steffen Udluft, Thomas Runkler

State-of-the-art reinforcement learning algorithms mostly rely on being allowed to directly interact with their environment to collect millions of observations.

Continuous Control Offline RL +1

Neural Topic Modeling with Continual Lifelong Learning

1 code implementation ICML 2020 Pankaj Gupta, Yatin Chaudhary, Thomas Runkler, Hinrich Schütze

To address the problem, we propose a lifelong learning framework for neural topic modeling that can continuously process streams of document collections, accumulate topics and guide future topic modeling tasks by knowledge transfer from several sources to better deal with the sparse data.

Data Augmentation Information Retrieval +1

Visual Evaluation of Generative Adversarial Networks for Time Series Data

no code implementations23 Dec 2019 Hiba Arnout, Johannes Kehrer, Johanna Bronner, Thomas Runkler

This is particularly true when parts of the training data have been artificially generated to overcome common training problems such as lack of data or imbalanced dataset.

Time Series

Lifelong Neural Topic Learning in Contextualized Autoregressive Topic Models of Language via Informative Transfers

no code implementations29 Sep 2019 Yatin Chaudhary, Pankaj Gupta, Thomas Runkler

in topic modeling, (2) A novel lifelong learning mechanism into neural topic modeling framework to demonstrate continuous learning in sequential document collections and minimizing catastrophic forgetting.

Data Augmentation Language Modelling +1

Neural Architectures for Fine-Grained Propaganda Detection in News

no code implementations WS 2019 Pankaj Gupta, Khushbu Saxena, Usama Yaseen, Thomas Runkler, Hinrich Schütze

To address the tasks of sentence (SLC) and fragment level (FLC) propaganda detection, we explore different neural architectures (e. g., CNN, LSTM-CRF and BERT) and extract linguistic (e. g., part-of-speech, named entity, readability, sentiment, emotion, etc.

Propaganda detection

Data Association with Gaussian Processes

no code implementations16 Oct 2018 Markus Kaiser, Clemens Otte, Thomas Runkler, Carl Henrik Ek

The data association problem is concerned with separating data coming from different generating processes, for example when data come from different data sources, contain significant noise, or exhibit multimodality.

Gaussian Processes Variational Inference

Neural Relation Extraction Within and Across Sentence Boundaries

1 code implementation11 Oct 2018 Pankaj Gupta, Subburam Rajaram, Hinrich Schütze, Bernt Andrassy, Thomas Runkler

iDepNN models the shortest and augmented dependency paths via recurrent and recursive neural networks to extract relationships within (intra-) and across (inter-) sentence boundaries.

Relation Extraction

Sensitivity Analysis for Predictive Uncertainty in Bayesian Neural Networks

no code implementations10 Dec 2017 Stefan Depeweg, José Miguel Hernández-Lobato, Steffen Udluft, Thomas Runkler

We derive a novel sensitivity analysis of input variables for predictive epistemic and aleatoric uncertainty.

Bayesian Alignments of Warped Multi-Output Gaussian Processes

no code implementations NeurIPS 2018 Markus Kaiser, Clemens Otte, Thomas Runkler, Carl Henrik Ek

We apply the method to the real-world problem of finding common structure in the sensor data of wind turbines introduced by the underlying latent and turbulent wind field.

Gaussian Processes Time Series

Cannot find the paper you are looking for? You can Submit a new open access paper.