AI Paper Atlas

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

Compositional Generalization through Gradient Search in Nonparametric Latent Space

Haruki Shirakami, James Henderson

ICLR 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Contrastive representation theory

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 KL-regularization (ablation)Abduction Transformer · 1-D ARC OOD compositionsolve rate (%): 16.7 -> 25.1Improved
add KL-regularization (ablation)Abduction Transformer · SRAVEN OOD composition (train on 1%)solve rate (%): 16.8 -> 46.1Improved
add KL-regularization (ablation)Abduction Transformer · SRAVEN OOD composition (train on 90%)solve rate (%): 33.7 -> 96.4Improved
add test-time gradient search (ablation)Abduction Transformer · 1-D ARC OOD compositionsolve rate (%): 0.1 -> 25.1Improved
add test-time gradient search (ablation)Abduction Transformer · SRAVEN OOD composition (train on 1%)solve rate (%): 20.9 -> 46.1Improved
add test-time gradient search (ablation)Abduction Transformer · SRAVEN OOD composition (train on 90%)solve rate (%): 96.7 -> 96.4No clear effect
use nonparametric mixture latent instead of single-vector latent (LPN)Abduction Transformer · 1-D ARC OOD compositionsolve rate (%): 1.9 -> 25.1Improved
use nonparametric mixture latent instead of single-vector latent (LPN)Abduction Transformer · SRAVEN OOD composition (train on 1%)solve rate (%): 37.1 -> 46.1Improved
use nonparametric mixture latent instead of single-vector latent (LPN)Abduction Transformer · SRAVEN OOD composition (train on 90%)solve rate (%): 93.5 -> 96.4Improved
stochastic Dirichlet-process encoder instead of deterministic encoder (ablation)Abduction Transformer · 1-D ARC OOD compositionsolve rate (%): 0.1 -> 25.1Improved
stochastic Dirichlet-process encoder instead of deterministic encoder (ablation)Abduction Transformer · SRAVEN OOD composition (train on 1%)solve rate (%): 10.8 -> 46.1Improved
stochastic Dirichlet-process encoder instead of deterministic encoder (ablation)Abduction Transformer · SRAVEN OOD composition (train on 90%)solve rate (%): 27.2 -> 96.4Improved

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