AI Paper Atlas

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

Controlling Neural Networks with Rule Representations

Sungyong Seo, Sercan O Arik, Jinsung Yoon, Xiang Zhang, Kihyuk Sohn, Tomas Pfister

NeurIPS 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Contrastive representation theory · 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 DeepCTRL with tuned alpha instead of fixed-lambda Task & Rule regularizationDeepCTRL · Double pendulum testrule verification ratio (%): 71.6 -> 80.2Improved
Apply DeepCTRL instead of TaskOnly on weaker-correlation Neg0.1 items (model from Neg0.2)DeepCTRL · Sales Neg0.1MAE: 0.05 -> 0.045Improved
Transfer model from source Neg0.2 to Neg0.1 target (TaskOnly degradation)TaskOnly · Sales Neg0.1MAE: 0.041 -> 0.05Hurt
Transfer model from source Neg0.2 to Neg0.1 target (DeepCTRL degradation)DeepCTRL · Sales Neg0.1MAE: 0.042 -> 0.045Hurt
Sample alpha from Beta(1,1) (uniform) instead of Beta(0.1,0.1) (ablation)DeepCTRL · Double pendulumMAE and verification ratioHurt
Vary Beta(beta,beta) prior around beta=0.1 (ablation)DeepCTRL · All datasetsMAENo clear effect
Replace stochastic coupling with Concat or Add of rule and data representations (ablation)DeepCTRL · Double pendulumMAE at intermediate alphaHurt
Concatenate alpha to input instead of two-passage architecture (ablation)DeepCTRL · Double pendulumtask error and verification ratioNo clear effect
Adapt rho every epoch instead of fixing it from initial losses (ablation)DeepCTRL · Double pendulumMAE and verification ratioNo clear effect
Increase perturbation upper bound u (ablation)DeepCTRL · Cardiovascularcross entropy / accuracy at alpha>0Hurt
Add rule via perturbation loss, compare TaskOnly vs DeepCTRL on Source cardiovascular dataDeepCTRL · Cardiovascular Sourcecross entropyHurt
Increase alpha up to 1.4 (extrapolation beyond training range)DeepCTRL · Cardiovascular Target 1cross entropy: 1.0 -> 1.4Improved

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