Timezone: »

 
Oral
Infinite Mixture Prototypes for Few-shot Learning
Kelsey Allen · Evan Shelhamer · Hanul Shin · Josh Tenenbaum

Wed Jun 12 05:00 PM -- 05:05 PM (PDT) @ Grand Ballroom
in Representation Learning »

We propose infinite mixture prototypes to adaptively represent both simple and complex data distributions for few-shot learning. Our infinite mixture prototypes represent each class by a set of clusters, unlike existing prototypical methods that represent each class by a single cluster. By infer-ring the number of clusters, infinite mixture prototypes interpolate between nearest neighbor and prototypical representations, which improves ac-curacy and robustness in the few-shot regime. We show the importance of adaptive capacity for capturing complex data distributions such as alpha-bets, with 25% absolute accuracy improvements over prototypical networks, while still maintain-ing or improving accuracy on the standard Omniglot and mini-ImageNet benchmarks. In clustering labeled and unlabeled data by the same clustering rule, infinite mixture prototypes achieves state-of-the-art semi-supervised accuracy. As a further capability, we show that infinite mixture prototypes can perform purely unsupervised clustering, unlike existing prototypical methods.

[ Slides]  [ Video] 

Author Information

Kelsey Allen (Massachusetts Institute of Technology)
Evan Shelhamer (UC Berkeley)
Hanul Shin (Massachusetts Institute of Technology)
Josh Tenenbaum (MIT)

Joshua Brett Tenenbaum is Professor of Cognitive Science and Computation at the Massachusetts Institute of Technology. He is known for contributions to mathematical psychology and Bayesian cognitive science. He previously taught at Stanford University, where he was the Wasow Visiting Fellow from October 2010 to January 2011. Tenenbaum received his undergraduate degree in physics from Yale University in 1993, and his Ph.D. from MIT in 1999. His work primarily focuses on analyzing probabilistic inference as the engine of human cognition and as a means to develop machine learning.

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors