pith. sign in

arxiv: 1702.08563 · v3 · pith:PCRXLYJYnew · submitted 2017-02-27 · 💻 cs.CL

Soft Label Memorization-Generalization for Natural Language Inference

classification 💻 cs.CL
keywords labelssofttraininglabeldatageneralizationnoisednns
0
0 comments X
read the original abstract

Often when multiple labels are obtained for a training example it is assumed that there is an element of noise that must be accounted for. It has been shown that this disagreement can be considered signal instead of noise. In this work we investigate using soft labels for training data to improve generalization in machine learning models. However, using soft labels for training Deep Neural Networks (DNNs) is not practical due to the costs involved in obtaining multiple labels for large data sets. We propose soft label memorization-generalization (SLMG), a fine-tuning approach to using soft labels for training DNNs. We assume that differences in labels provided by human annotators represent ambiguity about the true label instead of noise. Experiments with SLMG demonstrate improved generalization performance on the Natural Language Inference (NLI) task. Our experiments show that by injecting a small percentage of soft label training data (0.03% of training set size) we can improve generalization performance over several baselines.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Quantifying and Predicting Disagreement in Graded Human Ratings

    cs.CL 2026-05 unverdicted novelty 5.0

    Annotation disagreement on toxic language can be moderately predicted from textual features, with high-opposition items proving harder for models to estimate accurately.

  2. Modeling Human Perspectives with Socio-Demographic Representations

    cs.CL 2026-04 unverdicted novelty 5.0

    Socio-Contrastive Learning jointly learns socio-demographic representations and textual features via contrastive objectives to predict annotator perspectives more accurately than concatenation baselines.