1

13.09.2021 Meeting Notes

|

Meeting Details

Date : 13th September 2022

Time : 12:30 – 1:30

Location : Chair of CPS, Montanuniverität Leoben

Participants: Univ.-Prof. Dr. Elmar Rueckert, Vedant Dave

Agenda

  1. Learning Consistent Forward and Inverse Dynamics.

Topic 1: Idea Development

  1. Thinking in terms of Closed loop systems and feedback controllers.
  2. Regularizing Forward model via Inverse model.
  3. Single-step and Multi-step prediction models.
  4. Comparing Multi-step predictions with Movement Primitives.

Topic 2: Toy Example

  1. Generate a toy dataset(Temperature) with just single parameter.
  2. Try forward model to approximate out-of-distribution testing data.
  3. If it fails, try to regularize it with Inverse model and check if it works out.

Literature

Inverse Dynamic Predictions

  1. Cooper, Richard. (2010). Forward and Inverse Models in Motor Control and Cognitive Control. Proceedings of the International Symposium on AI Inspired Biology – A Symposium at the AISB 2010 Convention.
  2. Moore, Andrew. “Fast, robust adaptive control by learning only forward models.” Advances in neural information processing systems 4 (1991).

Next Meeting

TBD