Summary of The Reachability Problem For Neural-network Control Systems, by Christian Schilling and Martin Zimmermann
The Reachability Problem for Neural-Network Control Systems
by Christian Schilling, Martin Zimmermann
First submitted to arxiv on: 6 Jul 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computational Complexity (cs.CC); Logic in Computer Science (cs.LO); Systems and Control (eess.SY)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A feedforward neural network-based control system is developed to periodically compute a control input for a plant component. The reachability problem is investigated, which asks whether a set of target states can be reached from a given set of initial states. Despite the simplicity of the controller and plant, the authors show that the reachability problem becomes undecidable even with fixed-depth neural networks and trivial plants. However, they also demonstrate that the problem becomes semi-decidable when the plant and input/output sets are defined by automata over infinite words. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A control system is like a thermostat that helps keep something stable. In this case, we’re looking at a special kind of computer program that tries to reach certain goals from starting points. It might seem simple, but it turns out that making sure we can always reach those goals is actually impossible for some types of systems. However, if we give the system more information about what’s allowed and what’s not, then we can make it work again. |
Keywords
* Artificial intelligence * Neural network
