4 papers
Estimating the expected output of wide random MLPs more efficiently than sampling
Wilson Wu, Victor Lecomte, Michael Winer +3
By far the most common way to estimate an expected loss in machine learning is to draw samples, compute the loss on each one, and take the empirical average. However, sampling is n…
Bayesian Influence Functions for Hessian-Free Data Attribution
Philipp Alexander Kreer, Wilson Wu, Maxwell Adam +2
Classical influence functions face significant challenges when applied to deep neural networks, primarily due to non-invertible Hessians and high-dimensional parameter spaces. We p…
C*-like modules and matrix -operator norms
Alessandra Calin, Ian Cartwright, Luke Coffman +5
We present a generalization of Hölder duality to algebra-valued pairings via -modules. Hölder duality states that if and are conjugate expon…
Towards a unified and verified understanding of group-operation networks
Wilson Wu, Louis Jaburi, Jacob Drori +1
A recent line of work in mechanistic interpretability has focused on reverse-engineering the computation performed by neural networks trained on the binary operation of finite grou…