AI Paper Atlas

Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning

Meta-learning from Tasks with Heterogeneous Attribute Spaces

Tomoharu Iwata, Atsutoshi Kumagai

NeurIPS 2020 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Meta-augmentation and meta-objectives · 11 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 one round of attribute/response relation inference (L=0 -> L=1) (ablation)Ours · OpenML regression (17 target tasks)MSE: 0.834 -> 0.788Improved
increase iterations from L=1 to L=2 (ablation)Ours · OpenML regression (17 target tasks)MSE: 0.788 -> 0.777Improved
increase iterations beyond L=2 (ablation)Ours · OpenML regression (17 target tasks)MSE: 0.777 -> 0.781 (L=5)No clear effect
sample attributes and responses randomly during training (ablation)Ours · OpenML regression (17 target tasks)MSE: 0.848 -> 0.788Improved
use heterogeneous meta-learning inference network instead of Neural ProcessNP · OpenML regression (17 target tasks)MSE: 0.845 -> 0.788Improved
replace DS+MAML with OursDS+MAML · OpenML regression (17 target tasks)MSE: 0.854 -> 0.788Improved
fine-tune on support setDS · OpenML regression (17 target tasks)MSE: 0.896 -> 0.887No clear effect
use proposed classifier prototypes vs DSDS · OpenML classificationaccuracy: 0.621 -> 0.646Improved
increase number of training tasksOurs · OpenML regression (17 target tasks)MSEImproved
increase support set sizeOurs · OpenML regression (17 target tasks)MSEImproved

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