AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Should We Be Pre-training? An Argument for End-task Aware Training as an Alternative

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

ICLR 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 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
add the end-task objective through multi-task training with TAPT MLMRoBERTa_base continued pre-training · ACL-ARC, SCIERC, and CHEMPROTF1: ACL-ARC: 67.74_{3.68} -> 70.48_{4.42}; SCIERC: 79.53_{1.93} -> 80.81_{0.74}; CHEMPROT: 82.17_{0.65} -> 84.29_{0.63}Improved
meta-learn task weights for end-task and TAPT MLMTAPT continued pre-training · ACL-ARC, SCIERC, and CHEMPROTF1: ACL-ARC: 67.74_{3.68} -> 70.08_{4.70}; SCIERC: 79.53_{1.93} -> 81.48_{0.82}; CHEMPROT: 82.17_{0.65} -> 84.49_{0.50}Improved
add the end-task objective to DAPT and TAPT through multi-task trainingDAPT+TAPT · SCIERC and CHEMPROT, n = 10 × |Train|F1: SCIERC: 77.62_{1.38} -> 81.02_{1.24}; CHEMPROT: 78.22_{0.74} -> 83.77_{0.60}Improved
add the end-task objective to DAPT and TAPT, using fixed equal task weightsDAPT+TAPT · ACL-ARC, n = 10 × |Train|macro-F1: 69.12_{5.76} -> 71.58_{1.65}; p = 0.110Mixed
use META-TARTAN instead of MT-TARTAN to adapt weights for domain-only DAPTDAPT-only TARTAN · ACL-ARC, SCIERC, and CHEMPROTF1: ACL-ARC: 69.27_{0.96} -> 71.19_{4.88}; SCIERC: 81.53_{0.99} -> 82.08_{1.19}; CHEMPROT: 80.26_{3.79} -> 82.31_{0.75}Improved
use a separate meta-classification head instead of the end-task training head (ablation)META-TARTAN with ACL-ARC and DAPT · ACL-ARCF1: a delta of 1.7 F1 pointsImproved
use a separate meta-head rather than the same end-task head for calculating meta-gradients (ablation)META-TARTAN · ACL-ARC with DAPTauxiliary task weight utilizationImproved
use task-aware multi-task or meta-learning instead of task-agnostic pre-trainingCIFAR100 Medium-Sized Mammals classification · Medium-Sized Mammalsaccuracy: 46.7_{2.2} -> 51.3_{1.2}; 46.7_{2.2} -> 52.3_{3.8}Improved
use META-TARTAN instead of TAPTHYPER-PARTISAN classification · HYPER-PARTISANaccuracy: 93.39_{2.26} -> 96.84_{1.72}Improved
use META-TARTAN with 10 × |Train| DAPT data instead of full-data DAPTACL-ARC · ACL-ARCmacro-F1: 71.19_{4.88} vs 72.49_{3.28}Mixed
use META-TARTAN with 10 × |Train| auxiliary data instead of end-task-agnostic methods using full domain dataSCIERC · SCIERCmacro-F1: 81.41_{1.70} vs 79.97_{2.11}; 81.41_{1.70} vs 80.00_{1.08}Improved
use META-TARTAN with 10 × |Train| auxiliary data instead of full-data DAPT or DAPT+TAPTCHEMPROT · CHEMPROTmicro-F1: 83.38_{0.89} vs 86.54_{1.05}; 83.38_{0.89} vs 87.24_{0.81}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