Priberam Machine Learning Lunch Seminars - "Equivariant neural networks for recovery of Hadamard matrices", Augusto Peres (Inductiva Research Labs)

[Diogo Pernes]

Dear all,

 

The Priberam Machine Learning Seminars will continue to take place remotely via Zoom on Tuesdays at 1 PM.

Next Tuesday, April 5th, Augusto Peres, a researcher at Inductiva Research Labs, will present us his work about equivariant neural networks for recovery of Hadamard matrices at 1 PM (WEST) (zoom link: https://us02web.zoom.us/j/85100804324?pwd=UEZvMzhTbzQzWVppYzJaREUyeEJIdz09).

You can register for this event and keep watch on future seminars below:

https://www.eventbrite.pt/e/equivariant-neural-networks-for-recovery-of-hadamard-matrices-tickets-305867035537

We look forward to having you join us!

Kind regards,
Diogo Pernes

Priberam Labs
http://labs.priberam.com/

Priberam is hiring!

If you are interested in working with us please consult the available positions at priberam.com/careers. 

PRIBERAM SEMINARS   --  Zoom 896 1265 8540

__________________________________________________

Priberam Machine Learning Lunch Seminar
Speaker: Augusto Peres (Inductiva Research Labs)
Venue: https://us02web.zoom.us/j/85100804324?pwd=UEZvMzhTbzQzWVppYzJaREUyeEJIdz09

Date: Tuesday, April 5th, 2022
Time: 13:00 

Title:

Equivariant neural networks for recovery of Hadamard matrices

Abstract:

Geometric deep learning is an emerging area of research in machine learning focusing on exploiting symmetries in problems to improve models. Its goal is to understand how transformations to the input should affect the output and design neural networks around the corresponding inductive bias. We present a message passing neural network architecture designed to be equivariant to column and row permutations of a matrix. We illustrate its advantages over traditional architectures like multi-layer perceptrons (MLPs), convolutional neural networks (CNNs) and even Transformers, on the combinatorial optimization task of recovering a set of deleted entries of a Hadamard matrix. We argue that this is a powerful application of the principles of Geometric Deep Learning to fundamental mathematics, and a potential stepping stone toward more insights on the Hadamard conjecture using Machine Learning techniques.

Short Bio:

Augusto Peres is a researcher at Inductiva Research Labs. Currently, his main line of research is centered around machine learning for fundamental mathematics. More specifically, Augusto is focusing on the application of machine learning for solving combinatorial optimization problems. Augusto's other interests revolve around geometric deep learning, reinforcement learning, automata theory and formal methods.

More info: http://labs.priberam.com/Academia-Partnerships/Seminars.aspx

Eventbrite:

https://www.eventbrite.pt/e/equivariant-neural-networks-for-recovery-of-hadamard-matrices-tickets-305867035537