Adaptive Posterior Learning: few-shot learning with a surprise-based memory module

The ability to generalize quickly from few observations is crucial for intelligent systems. In this paper we introduce APL, an algorithm that approximates probability distributions by remembering the most surprising observations it has encountered... (read more)

PDF Abstract ICLR 2019 PDF ICLR 2019 Abstract
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Few-Shot Image Classification OMNIGLOT - 1-Shot, 1000 way APL Accuracy 68.9 # 1
Few-Shot Image Classification OMNIGLOT - 1-Shot, 20-way APL Accuracy 97.2% # 7
Few-Shot Image Classification OMNIGLOT - 1-Shot, 423 way APL Accuracy 73.5 # 1
Few-Shot Image Classification OMNIGLOT - 1-Shot, 5-way APL Accuracy 97.9 # 14
Few-Shot Image Classification OMNIGLOT - 5-Shot, 1000 way APL Accuracy 78.9 # 1
Few-Shot Image Classification OMNIGLOT - 5-Shot, 20-way APL Accuracy 97.6% # 16
Few-Shot Image Classification OMNIGLOT - 5-Shot, 423 way APL Accuracy 88 # 1
Few-Shot Image Classification OMNIGLOT - 5-Shot, 5-way APL Accuracy 99.9 # 2

Methods used in the Paper


METHOD TYPE
🤖 No Methods Found Help the community by adding them if they're not listed; e.g. Deep Residual Learning for Image Recognition uses ResNet