The Shape of Discovery: Detecting Topological Precursors to Technological Breakthroughs in the USPTO Patent Citation Network

Read full text Summarise with VIRAT
Authors: Christopher Ortiz
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Applies persistent homology to the U.S. patent citation network (~8M patents, ~118M citations, 1976-2023) to test whether topological signatures precede technological breakthroughs. The central finding is null: an apparent precursor signal (Wilcoxon p < 0.001, N=57 valid comparisons from 65-entry catalog) is entirely explained by a temporal confound — beta-1 (H1 feature counts) declines universally at ~1.8/year across all 28 CPC section pairs due to network maturation. Z-scores correlate almost perfectly with filing year (Spearman rho = -0.921, p < 1e-20). After per-pair linear detrending, all significance vanishes (Wilcoxon p = 0.85). The project demonstrates three findings: (1) the precursor hypothesis does not survive temporal detrending; (2) the patent knowledge landscape is systematically flattening over time (genuine structural change, confirmed after density control); (3) temporal trends in TDA on evolving networks produce spurious signals — a methodological cautionary tale. Seven confound robustness checks implemented. Conceptual design, analytical framework, and implementation by Claude (Opus 4.6, Anthropic) via claude.ai. Implementation by Claude Code. Facilitated by Christopher Ortiz.

Fields of Study & Subjects

Topological and Geometric Data Analysis Intellectual Property and Patents scientometrics and bibliometrics research

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
22 March 2026
Language
EN