AI Paper Atlas

Research map / Optimization / Implicit bias and generalization

Learning to Reason with Neural Networks: Generalization, Unseen Data and Boolean Measures

Emmanuel Abbe, Samy Bengio, Elisabetta Cornacchia, Jon Kleinberg, Aryo Lotfi, Maithra Raghu, Chiyuan Zhang

NeurIPS 2022 · Optimization › Implicit bias and generalization · approach: Stability-based generalization analysis · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
increase depth of linear neural network (holdout generalization error vs Boolean influence)linear neural network, hidden size 256 · linear target function on 11 bits, canonical holdoutgeneralization errorImproved
decrease initialization scale (alpha from 0.5 to 2.5) in 3-layer linear network3-layer linear neural network · linear target function on 11 bits, canonical holdoutgeneralization error: alpha in {0.5, 1, 1.5, 2, 2.5}Improved
use MLP-Mixer instead of MLP or Transformercanonical holdout PVR learning · PVR with 3 pointer bits, window size 3, majority votegap between generalization error and Boolean influenceHurt
use MLP or Transformer architecturecanonical holdout PVR learning · PVR with 3 pointer bits, window size 3, majority votegap between generalization error and Boolean influenceImproved
nonzero bias in the frozen linear target (hat f(empty) differs from hat f({k}))linear regression with GD · canonical holdout, frozen coordinate kexpected generalization errorHurt
minimum l2-norm interpolating solution instead of low-degree representation4-bit toy function with x2 frozen to 1 · canonical holdout examplegeneralization error: half the error of f_{-2}Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy