AI Paper Atlas

Research map / Learning theory and algorithms / Kernels and regression theory

The Pick-to-Learn Algorithm: Empowering Compression for Tight Generalization Bounds and Improved Post-training Performance

Dario Paccagnan, Marco Campi, Simone Garatti

NeurIPS 2023 · Learning theory and algorithms › Kernels and regression theory · approach: Information-theoretic generalization bounds · 3 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
use P2L with GD instead of PAC-BayesGD classifier · MNIST 1000 samples delta=0.035bound on risk: 0.177 -> 0.143Improved
use P2L with GD instead of PAC-BayesGD classifier · MNIST 1000 samples delta=0.035test misclassification: 0.088 -> 0.072Improved
use P2L with GD instead of test-set approachGD classifier · MNIST 1000 samples delta=0.035bound on risk: 0.129 -> 0.143Hurt
use P2L with GD instead of test-set approachGD classifier · MNIST 1000 samples delta=0.001bound on risk: 0.155 -> 0.163Hurt
use P2L with GD instead of test-set approachGD classifier · MNIST 1000 samples delta=0.035test misclassification: 0.079 -> 0.072Improved
use P2L instead of test-set approach (tightness)GD classifier · MNIST 1000 samples delta=0.035difference bound minus test risk: 0.050 -> 0.071Hurt
use P2L instead of PAC-Bayes (running time)GD classifier · MNISTrunning time: 4m 1s -> 2m 1sImproved
use P2L instead of test-set approach (running time)GD classifier · MNISTrunning time: 0m 5s -> 2m 1sHurt
vary pretrain portion of initial hypothesis h0 in P2L (ablation)P2L · MNISTtest misclassificationNo clear effect
use only 10% of data to pretrain h0 in P2L vs 60% prior for PAC-Bayes (ablation)P2L · MNISTbound and test performanceMixed
use P2L instead of test-set approachGD regressor · Synthetic regression N=200bound on risk: 0.250 -> 0.172Improved
use P2L instead of test-set approachGD regressor · Synthetic regression N=200test risk: 0.173 -> 0.064Improved

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