• Corpus ID: 244714184

Learning Physical Concepts in Cyber-Physical Systems: A Case Study

@article{Steude2021LearningPC,
  title={Learning Physical Concepts in Cyber-Physical Systems: A Case Study},
  author={Henrik S. Steude and Alexander Windmann and Oliver Niggemann},
  journal={ArXiv},
  year={2021},
  volume={abs/2111.14151}
}
Machine Learning (ML) has achieved great successes in recent decades, both in research and in practice. In Cyber-Physical Systems (CPS), ML can for example be used to optimize systems, to detect anomalies or to identify root causes of system failures. However, existing algorithms suffer from two major drawbacks: (i) They are hard to interpret by human experts. (ii) Transferring results from one systems to another (similar) system is often a challenge. Concept learning, or Representation… 

Figures and Tables from this paper

References

SHOWING 1-10 OF 41 REFERENCES

Data-driven anomaly detection in cyber-physical production systems

This paper outlines this data-driven approach to plant analysis using several use cases from industry and at the heart are intelligent anomaly detection approaches based on models of the system behaviors.

An Ensemble of Benchmarks for the Evaluation of AI Methods for Fault Handling in CPPS

An ensemble of simulated benchmark systems is presented, which allows for a broad evaluation of AI methods for fault handling and the suitability of the ensemble is shown using AI methodsfor fault handling such as anomaly detection, diagnosis and reconfiguration.

Discretization of hybrid CPPS data into timed automaton using restricted Boltzmann machines

A new dynamic predictive maintenance framework using deep learning for failure prognostics

A Deep Neural Network for Unsupervised Anomaly Detection and Diagnosis in Multivariate Time Series Data

This paper proposes a Multi-Scale Convolutional Recurrent Encoder-Decoder (MSCRED), to perform anomaly detection and diagnosis in multivariate time series data and demonstrates that MSCRED can outperform state-of-the-art baseline methods.

Discovering physical concepts with neural networks

This work models a neural network architecture after the human physical reasoning process, which has similarities to representation learning, and applies this method to toy examples to show that the network finds the physically relevant parameters, exploits conservation laws to make predictions, and can help to gain conceptual insights.

AI Feynman: A physics-inspired method for symbolic regression

This work develops a recursive multidimensional symbolic regression algorithm that combines neural network fitting with a suite of physics-inspired techniques and improves the state-of-the-art success rate.

Predictive Control Of Three-Tank-System Utilizing Both State-Space And Input-Output Models

The paper introduces a controller which integrates a predictive control synthesis based on a multivariable state-space model of a controlled system and an on-line identification of an ARX model

Modern Koopman Theory for Dynamical Systems

An overview of modern Koopman operator theory is provided, describing recent theoretical and algorithmic developments and highlighting these methods with a diverse range of applications, making it ideal for leveraging big-data and machine learning techniques.