Searching to Exploit Memorization Effect in Learning from Corrupted Labels

2019/11/06 by Quanming Yao, Hansi Yang, Yao, Quanming +6
Computer Science · #Advanced Multi-Objective Optimization Algorithms #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine Learning and Algorithms #Machine Learning and Data Classification

paper · pdf · doi:10.48550/arxiv.1911.02377

openalex publication_date 2019/11/06 · openalex created_date 2019/11/22 · openalex updated_date 2026/07/28

Abstract

Sample selection approaches are popular in robust learning from noisy labels. However, how to properly control the selection process so that deep networks can benefit from the memorization effect is a hard problem. In this paper, motivated by the success of automated machine learning (AutoML), we model this issue as a function approximation problem. Specifically, we design a domain-specific search space based on general patterns of the memorization effect and propose a novel Newton algorithm to solve the bi-level optimization problem efficiently. We further provide theoretical analysis of the algorithm, which ensures a good approximation to critical points. Experiments are performed on benchmark data sets. Results demonstrate that the proposed method is much better than the state-of-the-art noisy-label-learning approaches, and also much more efficient than existing AutoML algorithms.

Citations

• Natural Gradient Works Efficiently in Learning
• Proximal alternating linearized minimization for nonconvex and nonsmooth problems
• Practical Bayesian Optimization of Machine Learning Algorithms
• Convergence Rates of Inexact Proximal-Gradient Methods for Convex Optimization
• Designing Neural Network Architectures using Reinforcement Learning
• DARTS: Differentiable Architecture Search
• Adaptive Stochastic Natural Gradient Method for One-Shot Neural\n Architecture Search
• Understanding deep learning requires rethinking generalization
• Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results

Related

• MILD: Modeling the Instance Learning Dynamics for Learning with Noisy Labels
• Mitigating Memorization in Sample Selection for Learning with Noisy Labels
• Understanding Memorization from the Perspective of Optimization via Efficient Influence Estimation
• Mitigating Memorization of Noisy Labels via Regularization between Representations
• A Closer Look at Memorization in Deep Networks