Neumann, Klaus: Reliability of Extreme Learning Machines. 2014
Inhalt
- Abstract
- Introduction
- Random Projections and Extreme Learning Machines
- Extreme Learning Machine
- Improvements of the Extreme Learning Machine
- Review of Random Projection Methods
- Summary
- Robustness to Initializations by Batch Intrinsic Plasticity
- Reliability and Robustness to Initializations
- Intrinsic Plasticity
- Batch Intrinsic Plasticity: Methodology
- BIP and Single Neuron Behavior
- Experimental Results
- Conclusive Remarks
- Robustness to Drifts by Natural Intrinsic Plasticity
- Reliability and Drift Compensation
- Intrinsic Plasticity as Stochastic Gradient Descent
- The Natural Gradient for Intrinsic Plasticity
- Working-Point Transformation for IP
- Experimental Results
- Conclusive Remarks
- Reliability via Continuous Constraints
- Reliability via Continuous Constraints
- Related Work
- Embedding Discrete Constraints into ELMs
- From Discrete to Continuous Constraints
- Experimental Results
- Conclusive Remarks
- Reliable Control of the Bionic Handling Assistant
- Controlling the Bionic Handling Assistant
- Low-Level Control of the BHA
- Learning the BHA Data Set
- Models for the BHA Data Set
- Experimental Results on the BHA Data Set
- Experimental Results for Closed Loop Application
- Conclusive Remarks
- Stable Estimation of Dynamical Systems and Reliability
- Reliable Estimation of Dynamical Systems
- Neurally-Imprinted Stable Vector Fields
- What is a good Lyapunov Candidate for Learning?
- Experimental Results
- Conclusive Remarks
- Reliable Learning of the Ultrasonic Softening Effect
- Ultrasonic Softening with Application to Copper Bonding
- Copper Wire Deformation by Ultrasonic Softening
- Data-Driven Modeling with Integration of Prior Knowledge
- Experimental Results
- Conclusive Remarks
- Conclusion
- Appendix
- References