Local to Global Learning: Gradually Adding Classes for Training Deep Neural Networks

Citations Over TimeTop 12% of 2019 papers

Abstract

We propose a new learning paradigm, Local to Global Learning (LGL), for Deep Neural Networks (DNNs) to improve the performance of classification problems. The core of LGL is to learn a DNN model from fewer categories (local) to more categories (global) gradually within the entire training set. LGL is most related to the Self-Paced Learning (SPL) algorithm but its formulation is different from SPL. SPL trains its data from simple to complex, while LGL from local to global. In this paper, we incorporate the idea of LGL into the learning objective of DNNs and explain why LGL works better from an information-theoretic perspective. Experiments on the toy data, CIFAR-10, CIFAR-100, and ImageNet dataset show that LGL outperforms the baseline and SPL-based algorithms.

Related Papers

• → Communication to and in trains(2015)1 cited
• → Adversarial Robustness of Deep Learning: Theory, Algorithms, and Applications(2021)3 cited
• → Interpretable Mesomorphic Networks for Tabular Data(2023)1 cited
• EXPLORATION ON ORGANIZATION OF TRAIN OPERATION FOR URBAN RAIL TRANSIT(2001)
• Key Trains Identifying and its Reasonable Ratios in the High-Speed Rail Timetable(2015)