AI Science Tracker
EN

AI names colors much as humans do

Meta AI · 15 Mar 2021 · computer science

Discovery

Meta AI

AI names colors much as humans do

Source check: Source checked. Not independent scientific confirmation.

Finding: The trained systems approached the information-bottleneck accuracy/complexity frontier, and discrete versus continuous communication produced different efficiency patterns resembling human color naming. The paper reports 60 successful game runs across need levels, robustness settings and comparisons with 108 World Color Survey languages; this is a simulation and does not prove the origin of human language.

Technology: Deep-learning Speaker/Listener networks played a color-discrimination game over World Color Survey color chips; their learned naming systems were analyzed against human-language patterns.

Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.

AI contribution
Deep-learning Speaker/Listener networks played a color-discrimination game over World Color Survey color chips; their learned naming systems were analyzed against human-language patterns.
Test and result
The trained systems approached the information-bottleneck accuracy/complexity frontier, and discrete versus continuous communication produced different efficiency patterns resembling human color naming.
Limitation / still unknown
The paper reports 60 successful game runs across need levels, robustness settings and comparisons with 108 World Color Survey languages; this is a simulation and does not prove the origin of human language.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC8000426/

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.

Review date: 22 Nov 2026. Review scope: The primary PNAS full text reports a computational cognitive-science experiment in which trained Speaker/Listener neural networks developed efficient color-naming systems; title/DOI checks found no active or archived catalog match. The first located public version is the PNAS article, published online 2021-03-15; no earlier preprint was identified in the checked primary record..

No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.

Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Open in the full catalog for filters, charts and related discoveries. Browse computer science discoveries.

Related catalog entries

Links use documented relationships where available, plus shared catalog context. A shared topic or model family does not establish a scientific dependency.