Elena Burceanu

The best continual learning method for your task might depend more on *how much* of the model you fine-tune than *which* regularization strategy you use.

Seemingly innocuous choices about how to split a continuous data stream into discrete tasks can dramatically alter the conclusions of continual learning benchmarks, even before any model is trained.

Statistical feature processing and Learned Fourier features let deep learning finally challenge XGBoost's dominance on tabular data.