AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

Efficient Non-Parametric Optimizer Search for Diverse Tasks

Ruochen Wang, Yuanhao Xiong, Minhao Cheng, Cho-Jui Hsieh

NeurIPS 2022 · Probabilistic and scientific modeling › Molecules and proteins · approach: Prediction-powered and conformal inference · 0 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
replace random search with MCT traversal (ablation)MCT optimizer search · MNISTNETtest accuracy: 88.02% -> 89.43%Improved
replace random search with MCT traversal (ablation)MCT optimizer search · MNISTNETtraining loss (sum): 60.25 -> 59.25Improved
disable score thresholding (ablation)MCT optimizer search · MNISTNETcumulative training loss: 59.25 -> 60.25Hurt
search with MCT on super-tree instead of RL on hand-crafted patternsOptimizer search · CIFAR-10 ConvNettest accuracy: 76.21% -> 77.02%Improved
use 128 evaluations instead of over 10,000Optimizer search vs NOS-RL · CIFAR-10 ConvNetsearch budget (#evaluations): >10,000 -> 128Improved
use discovered optimizer instead of AdamGAT · OGBN-Productsaccuracy: 77.49% -> 80.15%Improved
use discovered optimizer instead of AdamGAT · Citeseeraccuracy: 71.70% -> 73.10%Improved
use discovered optimizer instead of AdamWBERT-base · MRPCaccuracy: 82.84 -> 86.64Improved
use discovered optimizer instead of AdamWBERT-base · RTEaccuracy: 65.97 -> 68.50Improved
use log-based discovered update instead of sign-based PGDPGD attack · CIFAR-10 Carmon2019attack success rate: 37.83% -> 38.35%Improved
compare discovered optimizer with Adaptive PGDAPGD · CIFAR-10 Gowal2020 (WRN-70-16)attack success rate: 32.00% -> 32.00%No clear effect
transfer a single discovered optimizer (products2) across graph datasetsDiscovered optimizer vs Adam · PubMedaccuracy: 78.20% -> 77.12%Hurt

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