54 citations · 65 across the 5 of their papers we have counts for
Showing cs.LGShow all
2 papers · 1 filter
cs.LG2026
What Should We Freeze? Guarded Freezing: Connectivity Shapes the Fine-Tuning of Pretrained Models
Leonel Aguilar
When adapting pre-trained models through fine-tuning, freezing weights alone might not preserve performance, as updates elsewhere can change the inputs to the frozen core, ultimate…
cs.LG2026
Does Your Neural Network Extrapolate? Feature Engineering as Identifiability Bias for OOD Generalization
Leonel Aguilar, Jan Nagler, Christoph Hoelscher +1
Successful deep neural networks discover salient features of data. We show when and why they fail to learn out-of-distribution (OOD)-relevant representations from an in-distributio…