Attentional Meta-learners for Few-shot Polythetic Classification

Ben J Day, Ramon Viñas Torné, Nikola Simidjievski, Pietro Lió
Proceedings of the 39th International Conference on Machine Learning, PMLR 162:4867-4889, 2022.

Abstract

Polythetic classifications, based on shared patterns of features that need neither be universal nor constant among members of a class, are common in the natural world and greatly outnumber monothetic classifications over a set of features. We show that threshold meta-learners, such as Prototypical Networks, require an embedding dimension that is exponential in the number of task-relevant features to emulate these functions. In contrast, attentional classifiers, such as Matching Networks, are polythetic by default and able to solve these problems with a linear embedding dimension. However, we find that in the presence of task-irrelevant features, inherent to meta-learning problems, attentional models are susceptible to misclassification. To address this challenge, we propose a self-attention feature-selection mechanism that adaptively dilutes non-discriminative features. We demonstrate the effectiveness of our approach in meta-learning Boolean functions, and synthetic and real-world few-shot learning tasks.

Cite this Paper

BibTeX
@InProceedings{pmlr-v162-day22a, title = {Attentional Meta-learners for Few-shot Polythetic Classification}, author = {Day, Ben J and Torn{\'e}, Ramon Vi{\~n}as and Simidjievski, Nikola and Li{\'o}, Pietro}, booktitle = {Proceedings of the 39th International Conference on Machine Learning}, pages = {4867--4889}, year = {2022}, editor = {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan}, volume = {162}, series = {Proceedings of Machine Learning Research}, month = {17--23 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v162/day22a/day22a.pdf}, url = {https://proceedings.mlr.press/v162/day22a.html}, abstract = {Polythetic classifications, based on shared patterns of features that need neither be universal nor constant among members of a class, are common in the natural world and greatly outnumber monothetic classifications over a set of features. We show that threshold meta-learners, such as Prototypical Networks, require an embedding dimension that is exponential in the number of task-relevant features to emulate these functions. In contrast, attentional classifiers, such as Matching Networks, are polythetic by default and able to solve these problems with a linear embedding dimension. However, we find that in the presence of task-irrelevant features, inherent to meta-learning problems, attentional models are susceptible to misclassification. To address this challenge, we propose a self-attention feature-selection mechanism that adaptively dilutes non-discriminative features. We demonstrate the effectiveness of our approach in meta-learning Boolean functions, and synthetic and real-world few-shot learning tasks.} }
Endnote
%0 Conference Paper %T Attentional Meta-learners for Few-shot Polythetic Classification %A Ben J Day %A Ramon Viñas Torné %A Nikola Simidjievski %A Pietro Lió %B Proceedings of the 39th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2022 %E Kamalika Chaudhuri %E Stefanie Jegelka %E Le Song %E Csaba Szepesvari %E Gang Niu %E Sivan Sabato %F pmlr-v162-day22a %I PMLR %P 4867--4889 %U https://proceedings.mlr.press/v162/day22a.html %V 162 %X Polythetic classifications, based on shared patterns of features that need neither be universal nor constant among members of a class, are common in the natural world and greatly outnumber monothetic classifications over a set of features. We show that threshold meta-learners, such as Prototypical Networks, require an embedding dimension that is exponential in the number of task-relevant features to emulate these functions. In contrast, attentional classifiers, such as Matching Networks, are polythetic by default and able to solve these problems with a linear embedding dimension. However, we find that in the presence of task-irrelevant features, inherent to meta-learning problems, attentional models are susceptible to misclassification. To address this challenge, we propose a self-attention feature-selection mechanism that adaptively dilutes non-discriminative features. We demonstrate the effectiveness of our approach in meta-learning Boolean functions, and synthetic and real-world few-shot learning tasks.
APA
Day, B.J., Torné, R.V., Simidjievski, N. & Lió, P.. (2022). Attentional Meta-learners for Few-shot Polythetic Classification. Proceedings of the 39th International Conference on Machine Learning, in Proceedings of Machine Learning Research 162:4867-4889 Available from https://proceedings.mlr.press/v162/day22a.html.

Related Material