Description

In this episode we interview Professor Matthew Turner, professor in the Cyber Physical Systems group at the university of Southampton, about his fascinating work on aircraft control, anti-windup, Lur’e systems, and his recent research into applying techniques from control theory to the analysis of neural networks.

The transcript for this episode is available at: https://docs.google.com/document/d/1r41Lcpv4MtxnqfMn8ZFfnAqiK5rlfcPZ19NQhfeqaFI/edit?usp=sharing  

Contact me at: lgh1g19@soton.ac.uk

We have an Instagram now!

References:

• PhD Thesis
• NationalResearch Council of Canada Aerospace Research Centre
• Nonlinear High Gain Feedback
• LMIs in Control
• AW for Model-Reference Adaptive Control
• AW for ILC
• Anti-Windup Review
• Two-Stage Anti-Windup
• Anti-Windup in Anticipation of Saturation (Zongli Lin)
• Matt Turner
• Chris Richards
• Aizerman Conjecture
• Ian Postlethwaite
• Jorge Sofrony
• Murray Kerr
• Oliver Brieger
• Circle and Popov Criterion
• Khalil Nonlinear Control Book
• Zames and Sakkary Gap Metric Paper
• George Zames
• Zames-Falb Multipliers Summary Article
• O’Shear
• Zames-Falb Multipliers: Don’t Panic
• Early Neural Network Work (with Guido Hermann)
• Additional NN Research
• Ross Drummond
• ReLU Network Analysis Using Lur’e Systems
• Antonis Papachristodolou
• Exploiting Sparsity in Neural Networks
• Robust Control Paper
• Malcolm Smith
• Glenn Vinnicombe

Other cool neural network/ control papers

• Synthesis of MPC Algorithms that Robustly Approximate Neural Networks
• Shrinking Down Neural Networks Using Inspiration from Control
• Designing Neural Networks that are More Robust