Extreme Classification via Adversarial Softmax Approximation

Abstract

Training a classifier over a large number of classes, known as 'extreme classification', has become a topic of major interest with applications in technology, science, and e-commerce. Traditional softmax regression induces a gradient cost proportional to the number of classes $C$, which often is prohibitively expensive. A popular scalable softmax approximation relies on uniform negative sampling, which suffers from slow convergence due a poor signal-to-noise ratio. In this paper, we propose a simple training method for drastically enhancing the gradient signal by drawing negative samples from an adversarial model that mimics the data distribution. Our contributions are three-fold: (i) an adversarial sampling mechanism that produces negative samples at a cost only logarithmic in $C$, thus still resulting in cheap gradient updates; (ii) a mathematical proof that this adversarial sampling minimizes the gradient variance while any bias due to non-uniform sampling can be removed; (iii) experimental results on large scale data sets that show a reduction of the training time by an order of magnitude relative to several competitive baselines.

Accepted for presentation at the Eighth International Conference on Learning Representations (ICLR 2020), https://openreview.net/forum?id=rJxe3xSYDS

References in corpus (2)

• Distributed Representations of Words and Phrases and their Compositionality
• One-vs-Each Approximation to Softmax for Scalable Estimation of Probabilities

Cited by in corpus (5)

• Semantic Models for the First-stage Retrieval: A Comprehensive Review
• Neural Transformation Learning for Deep Anomaly Detection Beyond Images
• Prototype Memory for Large-scale Face Representation Learning
• Unbiased Loss Functions for Extreme Classification With Missing Labels
• Disentangling Sampling and Labeling Bias for Learning in Large-Output Spaces