Title:
Deep Reinforcement Learning Using Data-Driven Reduced-Order Models Discovers and Stabilizes Low Dissipation Equilibria
Part 2: Data-driven dimension reduction, dynamic modeling, and control of complex chaotic systems
Deep Reinforcement Learning Using Data-Driven Reduced-Order Models Discovers and Stabilizes Low Dissipation Equilibria
Part 2: Data-driven dimension reduction, dynamic modeling, and control of complex chaotic systems
| dc.contributor.author | Graham, Michael | |
| dc.contributor.author | Zeng, Kevin | |
| dc.contributor.corporatename | Georgia Institute of Technology. School of Physics | en_US |
| dc.contributor.corporatename | University of Wisconsin-Madison. Dept. of Chemical and Biological Engineering | en_US |
| dc.date.accessioned | 2021-11-06T02:23:13Z | |
| dc.date.available | 2021-11-06T02:23:13Z | |
| dc.date.issued | 2021-10-27 | |
| dc.description | This is part two of the Nonlinear Science Webinar, titled "Data-driven dimension reduction, dynamic modeling, and control of complex chaotic systems", which was presented online via Bluejeans Meetings on October 27, 2021 at 3:00 p.m. | en_US |
| dc.description | Michael Graham is the Vilas Distinguished Achievement Professor and Harvey D. Spangler Professor in the Department of Chemical and Biological Engineering at the University of Wisconsin - Madison. He is the principal investigator of the Complex Flows and Fluids Research Group. The Graham group uses theory and computations to study problems in fluid dynamics, rheology and transport phenomena over a wide range of scales. | en_US |
| dc.description | Kevin Zeng is a graduate student at the University of Wisconsin - Madison and is a member of the Complex Flows and Fluids Research Group. | en_US |
| dc.description | Runtime: 37:52 minutes | en_US |
| dc.description.abstract | Mike Graham Overview: Our overall aim is to combine ideas from dynamical systems theory and machine learning to develop and apply reduced-order models of flow processes with complex chaotic dynamics. A particular aim is a minimal description of dynamics on manifolds of dimension much less than the nominal state dimension and use of these models to develop effective control strategies for reducing energy dissipation. | en_US |
| dc.description.abstract | Kevin Zeng: Deep reinforcement learning (RL), a data-driven method capable of discovering complex control strategies for high-dimensional systems, requires substantial interactions with the target system, making it costly when the system is computationally or experimentally expensive (e.g. flow control). We mitigate this challenge by combining dimension reduction via an autoencoder with a neural ODE framework to learn a low-dimensional dynamical model, which we substitute in place of the true system during RL training to efficiently estimate the control policy. We apply our method to data from the Kuramoto-Sivashinsky equation. With a goal of minimizing dissipation, we extract control policies from the model using RL and show that the model-based strategies perform well on the full dynamical system and highlight that the RL agent discovers and stabilizes a forced equilibrium solution, despite never having been given explicit information about this state’s existence. | |
| dc.format.extent | 37:52 minutes | |
| dc.identifier.uri | http://hdl.handle.net/1853/65416 | |
| dc.language.iso | en_US | en_US |
| dc.relation.ispartofseries | Nonlinear Science Seminar | |
| dc.subject | Dimension reduction | en_US |
| dc.subject | Kuramoto-Sivashinsky equation | en_US |
| dc.subject | Model based deep reinforcement learning | en_US |
| dc.title | Deep Reinforcement Learning Using Data-Driven Reduced-Order Models Discovers and Stabilizes Low Dissipation Equilibria | en_US |
| dc.title | Part 2: Data-driven dimension reduction, dynamic modeling, and control of complex chaotic systems | |
| dc.type | Moving Image | |
| dc.type.genre | Lecture | |
| dspace.entity.type | Publication | |
| local.contributor.corporatename | College of Sciences | |
| local.contributor.corporatename | School of Physics | |
| relation.isOrgUnitOfPublication | 85042be6-2d68-4e07-b384-e1f908fae48a | |
| relation.isOrgUnitOfPublication | 2ba39017-11f1-40f4-9bc5-66f17b8f1539 |
Files
Original bundle
1 - 4 of 4
No Thumbnail Available
- Name:
- zeng.mp4
- Size:
- 145.94 MB
- Format:
- MP4 Video file
- Description:
- Download video
No Thumbnail Available
- Name:
- zeng_videostream.html
- Size:
- 1.32 KB
- Format:
- Hypertext Markup Language
- Description:
- Streaming video
No Thumbnail Available
- Name:
- transcript.txt
- Size:
- 32.35 KB
- Format:
- Plain Text
- Description:
- Transcription
- Name:
- thumbnail.jpg
- Size:
- 60.58 KB
- Format:
- Joint Photographic Experts Group/JPEG File Interchange Format (JFIF)
- Description:
- Thumbnail
License bundle
1 - 1 of 1
No Thumbnail Available
- Name:
- license.txt
- Size:
- 3.13 KB
- Format:
- Item-specific license agreed upon to submission
- Description: