PyAerial is a Python implementation of the Aerial scalable neurosymbolic association rule miner for tabular data. It utilizes an under-complete denoising Autoencoder to learn a compact representation ...

As part of its commitment to improving the understanding of Classification within the Paralympic Movement, the IPC provides a free, open access online course. Whether you’re an Athlete, coach, ...

In this work, I explore methods for improving the cross-entropy loss alignment between SAE-reconstructed and original model logits. I find that: Rather than combining everything into one large and ...