Does the Data Induce Capacity Control in Deep Learning?
We show that the input correlation matrix of typical classification datasets has an eigenspectrum where, after a sharp initial drop, a large number of small eigenvalues are distributed uniformly over an exponentially large range. This structure is mirrored in a network trained on this data: we show that the Hessian and the Fisher Information Matrix (FIM) have eigenvalues that are spread uniformly over exponentially large ranges. We call such eigenspectra "sloppy" because sets of weights corresponding to small eigenvalues can be changed by large magnitudes without affecting the loss. Networks trained on atypical datasets with non-sloppy inputs do not share these traits and deep networks trained on such datasets generalize poorly. Inspired by this, we study the hypothesis that sloppiness of inputs aids generalization in deep networks. We show that if the Hessian is sloppy, we can compute non-vacuous PAC-Bayes generalization bounds analytically. By exploiting our empirical observation that training predominantly takes place in the non-sloppy subspace of the FIM, we develop data-distribution dependent PAC-Bayes priors that lead to accurate generalization bounds using numerical optimization.
Code (1)
Tasks
Deep LearningGeneralization BoundsSimilar Papers 제목 키워드 기반
Remove Symmetries to Control Model Expressivity
When symmetry is present in the loss function, the model is likely to be trapped in a low-capacity state that is sometimes known as a "collapse." Being trapped in these low-capacity states can be a major obstacle to trai…
modelTransplanting, inverting, and preventing a misalignment persona: method-conditional emergent misalignment in Qwen2.5
Emergent misalignment (EM) -- the broad misbehaviour a language model acquires after fine-tuning on narrow harmful data -- is mediated in Qwen2.5 models by a latent persona direction, and that direction is causal in open…
On the Impact of the Capacity Drop Phenomenon for Freeway Traffic Flow Control
Capacity drop is an empirically observed phenomenon in vehicular traffic flow on freeways whereby, after a critical density is reached, a state of congestion sets in, but the freeway does not become decongested again unt…
Sparsity, Superposition, and Forgetting: A Mechanistic Study of Representation Retention in Continual Learning
Continual learning (CL) systems often forget previously acquired knowledge, yet the mechanisms driving forgetting remain hard to isolate in practice because real datasets entangle many factors. We present a controlled, t…
Continual LearningDropout: Explicit Forms and Capacity Control
We investigate the capacity control provided by dropout in various machine learning problems. First, we study dropout for matrix completion, where it induces a data-dependent regularizer that, in expectation, equals the …
BIG-bench Machine LearningMatrix Completion