AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Feature Importance Ranking for Deep Learning

Maksymilian Wojtas, Ke Chen

NeurIPS 2020 · Learning theory and algorithms › Prediction and decision losses · approach: Active learning and subset selection

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 dual-net feature selection with a single classifier instead of best prior method with five classifiersdual-net · RNA-seqaccuracy: 0.9881 -> 0.9938Improved
keep ID feature in Glass datadual-net · Glassaccuracy: 75%-80% -> 90%+Improved
concatenate the mask with the masked features instead of using a placeholder value for removed features (ablation)operator net · feature selection tasksperformanceImproved
use weighted selector loss with weights 10, 5, 1selector net · feature selection taskssearch effectivenessImproved
compare with DFS for subset size above 15dual-net · Enhancer-PromoteraccuracyImproved
compare with LIME and SAN instance aggregationdual-net · Enhancer-Promoter, s 15 to 55accuracyImproved
compare with RFE and random forestdual-net · Enhancer-PromoteraccuracyImproved
use dual-net with alternate learning instead of RFEdual-net · Enhancer-Promotertraining time (s): 1,700 -> 1,100Improved
use dual-net instead of random forestdual-net · Enhancer-Promotertraining time (s): 2.5 -> 1,100Hurt
compare with CCM on TOX-171dual-net · TOX-171accuracyHurt
increase mask size s beyond 30dual-net · Yalefeature importance map quality: 30 best among 10,30,50,70,90Hurt
use operator validation loss with optimal mask only for early stoppingdual-net · synthetic and benchmark setsearly stopping signalImproved

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