scoring
A 0-10 rating of an AI paper's commercial viability. Higher scores cluster with code, demo, dataset, and beats-SOTA signals.
ScienceToStartup scores every new AI paper from 0 to 10 for commercial viability. The score is produced by an LLM under hard rules that link flag counts (has_code, has_demo, has_dataset, beats_sota) to numeric ranges, then sanity-checked against the corpus distribution. Higher scores are the primary signal that drives the Today Rail, the trends desk, and the Foresight ledger.
Source: curated glossary catalog. Freshness: git_versioned_curated_catalog.
Term API · apps/web/data/glossary/terms.ts