AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

AANG : Automating Auxiliary Learning

Lucio M. Dery, Paul Michel, Mikhail Khodak, Graham Neubig, Ameet Talwalkar

ICLR 2023 · Large language models › Language, safety and interpretability · approach: Data and evaluation methodology · 2 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 adaptive objective weighting rather than static weights (ablation)AANG-TD · Five NLP tasks, averagereported average performance: 81.07 -> 82.21Improved
Use 24 generated objectives rather than combining the 3 named objectivesAANG · Five NLP tasks, averagereported average performance: 81.55 -> 82.21Improved
Add external in-domain data and expand the search space from AANG-TD to AANG-TD+EDAANG-TD · ACL-ARCF1: 73.26 -> 73.70No clear effect
Increase sampled objectives per iteration from 3 of 24 to 6 of 24 (ablation)AANG-TD · ACL-ARCF1: 72.11 -> 73.26Improved
Increase sampled objectives per iteration from 3 of 24 to 6 of 24 (ablation)AANG-TD · SCIERCF1: 82.35 -> 82.98Improved
Increase sampled objectives per iteration from 3 of 24 to 6 of 24 (ablation)AANG-TD · SE-2016-6accuracy: 72.46 -> 72.46No clear effect
Increase sampled objectives per iteration from 3 of 24 to 6 of 24 (ablation)AANG-TD · CHEMPROTaccuracy: 83.91 -> 83.69Hurt
Increase sampled objectives per iteration from 3 of 24 to 6 of 24 (ablation)AANG-TD · H.PARTISANaccuracy: 98.46 -> 97.95Hurt
Use end-task-aware multitasking instead of end-task-agnostic auxiliary trainingAuxiliary learning on task data · Five NLP tasks, averagereported average performance: 78.72 -> 79.84Improved

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