Learning of Colors from Color Names: Distribution and Point Estimation

TL;DR

This paper investigates neural network models for estimating colors from color names, highlighting the effectiveness of simple sum of word embeddings and exploring both point and distribution estimates to account for subjective color descriptions.

Contribution

It compares various neural architectures and emphasizes the surprisingly strong performance of sum of word embeddings in color estimation tasks.

Findings

Sum of word embeddings outperforms more complex models.

Distribution estimates capture subjective color variations.

Neural models can effectively interpret color descriptions.

Abstract

Color names are often made up of multiple words. As a task in natural language understanding we investigate in depth the capacity of neural networks based on sums of word embeddings (SOWE), recurrence (LSTM and GRU based RNNs) and convolution (CNN), to estimate colors from sequences of terms. We consider both point and distribution estimates of color. We argue that the latter has a particular value as there is no clear agreement between people as to what a particular color describes -- different people have a different idea of what it means to be ``very dark orange'', for example. Surprisingly, despite it's simplicity, the sum of word embeddings generally performs the best on almost all evaluations.