AI Paper Atlas

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

Understanding Few-Shot Learning: Measuring Task Relatedness and Adaptation Difficulty via Attributes

Minyang Hu, Hong Chang, Zong Guo, Bingpeng Ma, Shiguang Shan, Xilin CHEN

NeurIPS 2023 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Fine-tuning strategies for few-shot · 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
train on 64 selected classes instead of all 100 (remove less-related heterogeneous classes) (ablation)MatchingNet · miniImageNet->CUB 5-shotaccuracy: 50.0 -> 51.7Improved
train on 64 selected classes instead of all 100 (remove less-related heterogeneous classes) (ablation)ProtoNet · miniImageNet->CUB 5-shotaccuracy: 53.7 -> 52.6Hurt
train on 64 selected classes instead of all 100 (remove less-related heterogeneous classes) (ablation)RelationNet · miniImageNet->CUB 5-shotaccuracy: 52.6 -> 50.3Hurt
train on 64 selected classes instead of all 100 (remove less-related heterogeneous classes) (ablation)MAML · miniImageNet->CUB 5-shotaccuracy: 51.6 -> 49.7Hurt
train on 64 selected classes instead of all 100 (remove less-related heterogeneous classes) (ablation)Baseline++ · miniImageNet->CUB 5-shotaccuracy: 52.4 -> 55.5Improved
add 4 extra labeled samples to hard novel tasks (distance > 0.29, ~5-7% of tasks), test-time interventionfive FSL methods (ProtoNet, RelationNet, MatchingNet, MAML, Baseline++) · miniImageNet->CUB 5-way 1-shot, intervened tasksaccuracyImproved
increase from 1-shot to 5-shotAPNet · CUB 5-wayaccuracy: approximately 10% at distance 0.14; around 15% at distance 0.20Improved
increase average task attribute distanceAPNet · CUB and SUN 5-way 1-shot and 5-shotaccuracyHurt
increase average task attribute distanceMatchingNet, ProtoNet, RelationNet, MAML, Baseline++ · CUB and SUNaccuracyHurt
increase distance computed with CLIP auto-annotated attributesfive FSL methods · miniImageNet->CUBaccuracyHurt
remove pattern, shape or texture attributes from distance computation (ablation)TAD with auto-annotated attributes · miniImageNet->CUB 5-shotdistance-accuracy trendMixed

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