2017/09/27 by Lyndon White, White, Lyndon, Roberto Togneri +5 · 1 citation
Computer Science · Mathematics · Psychology · #Advanced Image and Video Retrieval Techniques #Art #Artificial intelligence #Categorization, perception, and language #Computation and Language (cs.CL) #Computer science #Distribution (mathematics) #Economics #Estimation #FOS: Computer and information sciences #Geometry #Image Retrieval and Classification Techniques #Mathematical analysis #Mathematics #Point (geometry) #cs.CL
paper · pdf · doi:10.48550/arxiv.1709.09360
published in arXiv (Cornell University) (Cornell University) · Implementation available at https://github.com/oxinabox/ColoringNames.jl/
openalex publication_date 2017/09/27 · arxiv created 2020/01/10 · arxiv updated 2020/01/13 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
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.