Abstract
The future of production is smart and autonomous and so should the technologies that allow it. The application of machine learning (ML) in the manufacturing sector has been designated as one of the key players in the digital transformation. In general, the tendency of most ML application fields goes into more complex algorithms and deeper architectures. However, it is not the only available approach and often times not the most suitable for all use cases in the manufacturing industry where environments and products are highly standard. ML has the potential to solve complex problems by finding and learning relationships that humans cannot identify. Our hypothesis is that these relationships are invariant and exist inside specific tasks but also between a wide catalogue of similar, although not yet known, problems. An algorithm that finds and solves these invariances at a general and abstract level creates a general solution that can be directly deployed into new use cases leading to an increase in efficiency by reducing the required effort, time, and knowledge and allowing to leverage ML. Through three preliminary use cases, we demonstrate the potential of applying ML at a general and abstract level to increase the flexibility and reduce the model, task, and dataset dependency of ML solutions. This approach facilitates the conversion of automated environments into autonomous environments.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
References
Zhang, D., et al.: The ai index 2021 annual report. arXiv preprint arXiv:2103.06312 (2021)
Balakrishnan, T., Chui, M., Hall, B., Henke, N.: The state of ai in 2020. https://www.mckinsey.com/business-functions/mckinsey-analytics/our-insights/global-survey-the-state-of-ai-in-2020 (2020)
Dwivedi, Y.K., et al.: Artificial intelligence (ai): Multidisciplinary perspectives on emerging challenges, opportunities, and agenda for research, practice and policy. Int. J. Inf. Manage. 57, 101994 (2019)
Matei, I., de Kleer, J., Somarakis, C., Rai, R., Baras, J.S.: Interpretable machine learning models: a physics-based view. arXiv preprint arXiv:2003.10025 (2020)
Vanschoren, J.: Meta-learning: A survey. arXiv preprint arXiv:1810.03548 (2018)
Zhuang, H., et al.: Interpretable learning-to-rank with generalized additive models. arXiv preprint arXiv:2005.02553 ()2020
Ng, A.: Ai transformation playbook. https://landing.ai/ai-transformation-playbook/ (2019)
Koch, G., Zemel, R., Salakhutdinov, R.: Siamese neural networks for one-shot image recognition. In: ICML Deep Learning Workshop, vol. 2, Lille (2015)
Zhou, Al., Knowles, T., Finn, C.: Meta-learning symmetries by reparameterization. arXiv preprint arXiv:2007.02933 (2020)
Fleming, N.: How artificial intelligence is changing drug discovery. Nature 557(7706), S55–S55 (2018)
Kleff, S., Meduri, A., Budhiraja, R., Mansard, N., Righetti, L.: High-frequency nonlinear model predictive control of a manipulator. In: ICRA 2021 (2021)
Saveriano, M., Abu-Dakka, F.J., Kramberger, A., Peternel, L.: Dynamic movement primitives in robotics: A tutorial survey. arXiv preprint arXiv:2102.03861 (2021)
Ratliff, N.D., Issac, J., Kappler, D., Birchfield, S., Fox, D.: Riemannian motion policies. arXiv preprint arXiv:1801.02854 (2018)
Tobin, J., Fong, R., Ray, A., Schneider, J., Zaremba, W., Abbeel, P.: Domain randomization for transferring deep neural networks from simulation to the real world. In: 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 23–30. IEEE (2017)
Liang, J., Makoviychuk, V., Handa, A., Chentanez, N., Macklin, M., Fox, D.: Gpu-accelerated robotic simulation for distributed reinforcement learning. In: Conference on Robot Learning, pp. 270–282. PMLR (2018)
Acknowledgments
We would like to thank ARENA2036 e.V., our colleagues Felix Euteneuer and Priyadarshini Krishna Shobha from Mercedes-Benz AG, and Dr. Marc Gebauer and Marlon Lehmann form BTU Cottbus-Senftenberg for their support during this research.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 Springer Nature Switzerland AG
About this paper
Cite this paper
Daniel, D.Á., Lars, V., Ulrich, B. (2022). Clustered Problems and Machine Learning Methodologies: A New Approach. In: Andersen, AL., et al. Towards Sustainable Customization: Bridging Smart Products and Manufacturing Systems. CARV MCPC 2021 2021. Lecture Notes in Mechanical Engineering. Springer, Cham. https://doi.org/10.1007/978-3-030-90700-6_50
Download citation
DOI: https://doi.org/10.1007/978-3-030-90700-6_50
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-90699-3
Online ISBN: 978-3-030-90700-6
eBook Packages: EngineeringEngineering (R0)