Deep k-NN for Noisy Labels

04/26/2020
by   Dara Bahri, et al.
0

Modern machine learning models are often trained on examples with noisy labels that hurt performance and are hard to identify. In this paper, we provide an empirical study showing that a simple k-nearest neighbor-based filtering approach on the logit layer of a preliminary model can remove mislabeled training data and produce more accurate models than many recently proposed methods. We also provide new statistical guarantees into its efficacy.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
11/09/2020

A Survey of Label-noise Representation Learning: Past, Present and Future

Classical machine learning implicitly assumes that labels of the trainin...
research
08/14/2016

Bayesian Model Selection Methods for Mutual and Symmetric k-Nearest Neighbor Classification

The k-nearest neighbor classification method (k-NNC) is one of the simpl...
research
07/20/2023

Differences Between Hard and Noisy-labeled Samples: An Empirical Study

Extracting noisy or incorrectly labeled samples from a labeled dataset w...
research
08/27/2023

Label Denoising through Cross-Model Agreement

Learning from corrupted labels is very common in real-world machine-lear...
research
05/09/2023

On the Information Capacity of Nearest Neighbor Representations

The von Neumann Computer Architecture has a distinction between computat...
research
09/02/2019

An Adjusted Nearest Neighbor Algorithm Maximizing the F-Measure from Imbalanced Data

In this paper, we address the challenging problem of learning from imbal...
research
06/12/2018

Learning to Automatically Generate Fill-In-The-Blank Quizzes

In this paper we formalize the problem automatic fill-in-the-blank quest...

Please sign up or login with your details

Forgot password? Click here to reset