Researchllmstrokepatentsinnovation metrics
LLM Identifies Stroke Innovation Patent Trends
|
8.3
Marcus et al. (2026) analyze patent and publication records from 1993–2023 to quantify innovation in stroke care, using a 13-billion-parameter Llama LLM to filter patents. The LLM achieved 99.2% accuracy (96.5% sensitivity, 99.6% specificity), reducing 237,035 patents to 28,225 stroke-related filings grouped into seven innovation clusters. Pharmacological treatments dominated earlier patents but plateaued, while AI methods, rehabilitation devices, and imaging show rapid patent growth.
Scoring Rationale
Strong methodology and peer-reviewed evidence supports high practical relevance, but scope limited to stroke-specific innovation.
Practice with real Logistics & Shipping data
90 SQL & Python problems · 15 industry datasets
Used by DS/ML engineers at top companies
High-Value Overnight OrdersEasyDelivered International ShipmentsMediumOn-Time Delivery Rate by CarrierHard
250 free problems · No credit card
See all Logistics & Shipping problems
