AI Paper Atlas

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

Modular Meta-Learning with Shrinkage

Yutian Chen, Abram L. Friesen, Feryal Behbahani, Arnaud Doucet, David Budden, Matthew Hoffman, Nando de Freitas

NeurIPS 2020 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Meta-augmentation and meta-objectives · 5 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 learned shrinkage prior (σ-iMAML vs iMAML)iMAML · Large-data augmented Omniglottest accuracy (%): 72.8 -> 73.6Improved
add learned shrinkage prior (σ-Reptile vs Reptile)Reptile · Large-data augmented Omniglottest accuracy (%): 77.8 -> 78.9Improved
add shrinkage prior when few training instances per alphabetiMAML and Reptile · Small-data augmented Omniglottest accuracyImproved
add shrinkage prior to Reptile with 4 mins speech dataReptile · TTS naturalness, 4 minsMOS: 1.93 -> 1.98Improved
add shrinkage prior to Reptile with 8 mins speech dataReptile · TTS naturalness, 8 minsMOS: 2.09 -> 2.28Improved
adapt only 12 of 123 modules selected by σ² instead of all modules (ablation)Reptile · TTS voice similarity and naturalnessMOS: 2.09 -> 2.28Improved
adapt only the shrinkage-identified output layer rather than all layers (ablation)σ-iMAML and σ-Reptile · Large-data augmented Omniglottest accuracyNo clear effect
adapt penultimate layer only (ablation)σ-MAML-style short-adaptation image classification · standard Omniglot / miniImageNettest accuracyMixed
adapt only the first layer (ablation)shrinkage method · Sinusoid regressionerrorImproved
adapt all modules (SEA-All) instead of only speaker embedding (SEA-Emb) at 4 minsSEA · TTS naturalness, 4 minsMOS: 1.51 -> 1.41Hurt
adapt all modules (SEA-All) instead of only speaker embedding (SEA-Emb) at 8 minsSEA · TTS naturalness, 8 minsMOS: 1.57 -> 1.73Improved
meta-train with Reptile instead of finetuning SEA-AllWaveNet TTS · TTS naturalness, 8 minsMOS: 1.73 -> 2.09Improved

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