Equivariant Networks - Applications

Speaker

Anna Scaife, U. of Manchester

An introduction to equivariant neural networks, exploring the theoretical foundations of how symmetry constraints can be built directly into neural network architectures for more efficient and physically meaningful learning.

Topics Covered

• Symmetries and invariances in physics and astronomy
• Group theory foundations for neural networks
• Equivariant layers and architectures
• Translation, rotation, and permutation equivariance
• Applications to astronomical data analysis

Meeting Connection

Join the Meeting