Using predefined vector systems to speed up neural network multimillion class classification

TL;DR

This paper introduces a method to significantly accelerate neural network label prediction by leveraging known latent space geometry and vector systems, reducing complexity from O(n) to O(1).

Contribution

It presents a novel approach that uses predefined vector systems to speed up classification in neural networks without affecting accuracy.

Findings

Achieves up to 11.6 times faster inference

Reduces label prediction complexity to O(1)

Maintains training accuracy while accelerating inference

Abstract

Label prediction in neural networks (NNs) has O(n) complexity proportional to the number of classes. This holds true for classification using fully connected layers and cosine similarity with some set of class prototypes. In this paper we show that if NN latent space (LS) geometry is known and possesses specific properties, label prediction complexity can be significantly reduced. This is achieved by associating label prediction with the O(1) complexity closest cluster center search in a vector system used as target for latent space configuration (LSC). The proposed method only requires finding indexes of several largest and lowest values in the embedding vector making it extremely computationally efficient. We show that the proposed method does not change NN training accuracy computational results. We also measure the time required by different computational stages of NN inference and label prediction on multiple datasets. The experiments show that the proposed method allows to achieve up to 11.6 times overall acceleration over conventional methods. Furthermore, the proposed method has unique properties which allow to predict the existence of new classes.